CN111374712B - A kind of ultrasonic imaging method and ultrasonic imaging equipment - Google Patents

Abstract

The embodiment of the present application discloses an ultrasonic imaging method, which is applied to an ultrasonic imaging device, and the method includes: acquiring three-dimensional volume data of a first tissue to be measured; identifying a fetus from the three-dimensional volume data of the first tissue to be measured Three-dimensional volume data of the spine and three-dimensional volume data of the fetal ribs; rendering the three-dimensional volume data of the fetal spine and the three-dimensional volume data of the fetal ribs to obtain a three-dimensional ultrasound image of the fetal rib structure; The fetal ribs are marked in the three-dimensional ultrasonic image; and the marked three-dimensional ultrasonic image of the fetal rib structure is output. The embodiment of the present application discloses an ultrasonic imaging device.

Description

A kind of ultrasonic imaging method and ultrasonic imaging equipment

technical field

The embodiment of the present application relates to medical diagnosis technology, and relates to but not limited to an ultrasonic imaging method and an ultrasonic imaging device.

Background technique

Prenatal ultrasound examination is one of the most important items that mothers-to-be must check during pregnancy. Its main functions include determining the age of the fetus, analyzing the development of the fetus, detecting fetal deformities or abnormalities, taking photos and dynamic videos of the fetus, etc. Among them, the abnormal number of fetal ribs and the position of the end of the conus medullaris are one of the key items for prenatal ultrasound screening.

Abnormal number (or lesions) of fetal ribs (or vertebrae) is usually associated with fetal chromosomal abnormalities and some clinical syndromes, such as trisomy syndrome, thoracic dysplasia syndrome, and type I bone dysplasia. During the fetal period, the position of the end of the conus medullaris is of great significance for early screening of tethered cord syndrome. Because the position of the end of the conus medullaris is not fixed and changes with the gestational age, the spinal cord is stretched longitudinally due to congenital reasons, and the position of the end of the spinal cord decreases when pathological changes occur in the spinal cord. The resulting nerve damage syndrome is called spinal cord embolism. Department of syndrome. Failure to discover these congenital anomalies in the prenatal examination will bring a huge mental and economic burden to the patient's family and society, and even cause the patient's death in the neonatal period due to dyspnea caused by thoracic deformity.

Ultrasound is a safe, convenient, non-invasive and highly reproducible imaging technology. Two-dimensional and three-dimensional ultrasound can characteristically reflect the abnormality of fetal ribs, fetal spine and vertebral body, and has become the first choice for doctors to diagnose abnormal fetal ribs. . Among them, two-dimensional ultrasound can examine the fetal ribs and spinal cord from the sagittal plane, transverse plane and coronal plane, and can fully understand the situation of fetal ribs, fetal spine and spinal cord; while three-dimensional ultrasound can overcome fetal ribs and spinal cord examination. Two-dimensional ultrasound has problems such as lack of sense of space, poor fidelity, and difficult positioning. However, at this stage, doctors still need a lot of purely manual operations when using 3D ultrasound to examine fetal ribs and spinal cord, and the examination still has the following pain points:

The fetus may be in various positions during the examination, and the mid-sagittal plane of the fetal spine needs to be taken as the starting plane when acquiring a three-dimensional ultrasound volume. After obtaining the 3D data of fetal ribs and fetal spine, doctors need to have a very deep understanding of 3D space, so that they can select regions of interest through multiple manual rotation and translation geometric operations and virtual display (Virtual Reality, VR) under 3D ultrasound (VOI) cropping operation to determine the exact position of the missing and diseased ribs of the fetus or to determine the exact position of the conus medullaris.

Contents of the invention

In view of this, an embodiment of the present application provides an ultrasonic imaging method and an ultrasonic imaging device.

The technical scheme of the embodiment of the application is realized in this way:

On the one hand, an embodiment of the present application provides an ultrasonic imaging method applied to an ultrasonic imaging device, the method comprising:

Acquiring the three-dimensional volume data of the first tissue to be measured; identifying the three-dimensional volume data of the fetal spine and the three-dimensional volume data of the fetal ribs from the three-dimensional volume data of the first tissue to be measured; Rendering the three-dimensional volume data of the fetal ribs to obtain a three-dimensional ultrasound image of the fetal rib structure; marking the fetal ribs in the three-dimensional ultrasound image of the fetal rib structure and/or according to the identified three-dimensional volume data of the fetal ribs Determining the number of fetal ribs; outputting the labeled three-dimensional ultrasound image of the fetal rib structure and/or displaying the determined number of fetal ribs.

On the one hand, an embodiment of the present application provides an ultrasonic imaging method applied to an ultrasonic imaging device, the method comprising: acquiring three-dimensional volume data of a second tissue to be measured; obtaining three-dimensional volume data of the second tissue to be measured Identifying the three-dimensional volume data of the conus medullaris and the three-dimensional volume data of the lumbar spine; rendering the three-dimensional volume data of the conus medullaris and the three-dimensional volume data of the lumbar spine to obtain a three-dimensional ultrasonic image of the vertebral structure; The conus medullaris is marked in the ultrasound image; and the marked three-dimensional ultrasound image of the vertebral structure is output.

On the one hand, an embodiment of the present application provides an ultrasonic imaging method, which is applied to an ultrasonic imaging device, and the method includes: obtaining three-dimensional volume data of the fetus; based on the characteristics of the fetal ribs, identifying Three-dimensional volume data of fetal ribs; according to the identified three-dimensional volume data of fetal ribs, at least two ribs in the three-dimensional volume data of fetal ribs and multiple fetal ribs in the three-dimensional volume data of fetal ribs are obtained arranging the first plane or the first curved surface parallel or coincident with each other and/or obtaining an arrangement passing through at least one rib in the three-dimensional volume data of fetal ribs and aligning with multiple fetal ribs in the three-dimensional volume data of fetal ribs A second plane or a second curved surface where the surfaces intersect; according to the identified three-dimensional volume data of fetal ribs, obtain an image on the first plane or the first curved surface and/or obtain an image on the second plane or the second curved surface ; displaying the image on the first plane or the first curved surface as a two-dimensional image and/or displaying the image on the second plane or the second curved surface as a two-dimensional image.

On the one hand, an embodiment of the present application provides an ultrasonic imaging method applied to an ultrasonic imaging device, the method comprising: acquiring fetal three-dimensional volume data; based on the characteristics of the fetal conus medullaris, obtaining The conus medullaris area is identified in the data; the position of the conus medullaris area is determined according to the identified conus medullaris area; and the position of the conus medullaris area is displayed.

On the one hand, an embodiment of the present application provides an ultrasonic imaging device, including:

probe;

A transmitting circuit that excites the probe to transmit ultrasonic waves to the first tissue to be measured;

a receiving circuit for receiving, through the probe, ultrasonic echoes returned from the first tissue to be tested to obtain ultrasonic echo signals;

a processor, processing the ultrasonic echo signal to obtain a three-dimensional ultrasonic image of the marked fetal rib structure;

a monitor showing a three-dimensional ultrasound image of the marked fetal rib structure;

Wherein, the processor also performs the following steps:

Acquiring three-dimensional volume data of the first tissue to be measured according to the ultrasonic echo information;

identifying the three-dimensional volume data of the fetal spine and the three-dimensional volume data of the fetal ribs from the three-dimensional volume data of the first tissue to be measured;

Rendering the three-dimensional volume data of the fetal spine and the three-dimensional volume data of the fetal ribs to obtain a three-dimensional ultrasonic image of the fetal rib structure;

marking fetal ribs in the three-dimensional ultrasound image of the fetal rib structure and/or determining the number of fetal ribs from the identified three-dimensional volume data of the fetal ribs;

Outputting the labeled three-dimensional ultrasound image of the fetal rib structure and/or displaying the determined number of fetal ribs.

On the one hand, an embodiment of the present application provides an ultrasonic imaging device, including:

probe;

A transmitting circuit that excites the probe to transmit ultrasonic waves to the second tissue to be measured;

a receiving circuit for receiving ultrasonic echoes returned from the second tissue to be measured through the probe to obtain ultrasonic echo signals;

a processor, processing the ultrasound echo signal to obtain a three-dimensional ultrasound image of the marked vertebral structure;

a monitor showing a three-dimensional ultrasound image of the marked vertebral structure;

Wherein, the processor also performs the following steps:

Acquiring three-dimensional volume data of the second tissue to be measured according to the ultrasonic echo information;

identifying the three-dimensional volume data of the conus medullaris and the three-dimensional volume data of the lumbar spine from the second three-dimensional volume data;

Rendering the three-dimensional volume data of the conus medullaris and the three-dimensional volume data of the lumbar spine to obtain a three-dimensional ultrasound image of the vertebral structure;

marking said conus medullaris in a three-dimensional ultrasound image of said vertebral structure;

Outputting the marked three-dimensional ultrasonic image of the vertebral structure.

In the embodiment of the present application, the three-dimensional data of the fetal spine and the three-dimensional data of the fetal ribs are identified from the three-dimensional data of the first tissue to be measured, or the three-dimensional data of the conus medullaris and the lumbar spine are identified from the three-dimensional data of the second tissue to be measured. 3D data, render the identified 3D data, obtain the 3D ultrasound image of the fetal rib structure or the 3D ultrasound image of the spine, and mark the fetal rib or conus medullaris in the displayed 3D ultrasound image; thus, from the 3D volume Identify and detect the position of fetal ribs or conus medullaris in the data, automatically count the number of fetal ribs and calculate the relative position of conus medullaris and lumbar spine, simplify the workflow of fetal ribs or conus medullaris position inspection, improve the inspection efficiency, and allow doctors to It is liberated from complex and time-consuming operations, and reduces the dependence of fetal rib or conus medullaris position examination on the examiner's technique, improving examination efficiency.

Description of drawings

Fig. 1 is a structural schematic diagram 1 of an ultrasonic imaging device provided by an embodiment of the present application;

FIG. 2 is a schematic flow chart of the ultrasonic imaging method provided in the embodiment of the present application;

Fig. 3 is a schematic flow diagram II of the ultrasonic imaging method provided by the embodiment of the present application;

FIG. 4 is a schematic flow diagram III of the ultrasonic imaging method provided in the embodiment of the present application;

FIG. 5 is a schematic flow diagram IV of the ultrasonic imaging method provided by the embodiment of the present application;

Fig. 6 is a schematic flow diagram five of the ultrasonic imaging method provided by the embodiment of the present application;

Figure 7 is a schematic diagram of fetal ribs;

Fig. 8 is a schematic diagram of manually drawing anatomical trajectories in the related art;

Figure 9 is a schematic diagram of the conus medullaris;

FIG. 10 is a second schematic diagram of the composition and structure of the ultrasonic imaging device provided by the embodiment of the present application;

11 is a schematic diagram of the sequence of fetal ribs in the fetal rib structure;

Figure 12 is a schematic diagram of the straightening direction of the fetal ribs and spine in the fetal rib structure in the embodiment of the present application;

Fig. 13 is a schematic diagram showing cross-sections of all ribs in the embodiment of the present application;

Figure 14 is a schematic diagram showing the cross-section of the specified rib in the embodiment of the present application;

Figure 15 is a schematic diagram of the three-dimensional skeleton effect of the fetal rib structure in the embodiment of the present application;

Fig. 16 is a schematic diagram of the labeling effect of the conus medullaris in the VR image in the embodiment of the present application.

Detailed ways

The present application will be described in further detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the embodiments provided here are only used to explain the present application, and are not configured to limit the present application. In addition, the embodiments provided below are configured to implement some of the embodiments of the present application, rather than providing all the embodiments of the present application. In the case of no conflict, the technical solutions described in the embodiments of the present application can be combined in any manner implement.

FIG. 1 is a schematic structural block diagram of an ultrasonic imaging device 10 in an embodiment of the present application. The ultrasonic imaging device 10 may include a probe 100 , a transmitting circuit 101 , a transmitting/receiving selection switch 102 , a receiving circuit 103 , a beam forming circuit 104 , a processor 105 and a display 106 . The transmitting circuit 101 can stimulate the probe 100 to transmit ultrasonic waves to the target object. The receiving circuit 103 can receive the ultrasonic echo returned from the target object through the probe 100, so as to obtain the ultrasonic echo signal. The ultrasonic echo signal is sent to the processor 105 after undergoing beamforming processing by the beamforming circuit 104 . The processor 105 processes the ultrasound echo signal to obtain an ultrasound image of the target object. Ultrasound images obtained by processor 105 may be stored in memory 107 . These ultrasound images may be displayed on the display 106 .

Wherein, the target object includes at least one of the first tissue to be measured and the second tissue to be measured. The first tissue to be tested includes the fetal rib structure, and the second tissue to be tested includes the spine.

An embodiment of the present application provides an ultrasonic imaging method, which is applied to the ultrasonic imaging device shown in FIG. 1, as shown in FIG. 2, and the method includes:

S201. Acquiring three-dimensional volume data of the first tissue to be measured;

The first tissue to be tested includes: fetal rib structure and tissues other than fetal rib structure, such as amniotic fluid region, placenta, uterine wall and other tissues.

The doctor can scan the pregnant woman through the probe to obtain the three-dimensional volume data of the first tissue to be measured.

Here, the fetal rib structures belong to the hyperechoic region and appear as high gray levels in the ultrasound image.

S202. Identify the three-dimensional volume data of the fetal spine and the three-dimensional volume data of the fetal ribs from the three-dimensional volume data of the first tissue to be measured;

After obtaining the three-dimensional volume data of the first tissue to be measured based on S201, the three-dimensional volume data of the fetal spine and the three-dimensional volume data of the fetal ribs are identified from the acquired three-dimensional volume data. Here, identifying the three-dimensional volume data of the fetal spine and the three-dimensional volume data of the fetal ribs includes at least one of the following two identification methods:

Identification method 1. Firstly, the whole fetal rib structure is used as the identification object to identify the three-dimensional volume data of the fetal rib structure from the three-dimensional volume data of the first tissue to be measured, and then identify the fetal rib structure data from the three-dimensional volume data of the fetal rib structure. 3D volumetric data of fetal ribs.

The second identification method is to directly use the fetal rib as the identification object to identify the three-dimensional volume data of the fetal rib from the three-dimensional volume data of the first tissue to be measured.

Wherein, in identification mode one, the three-dimensional volume data of the fetal rib structure is identified from the three-dimensional volume data of the first tissue to be measured; The three-dimensional volume data of the tissue is segmented; the three-dimensional volume data of the fetal spine and the three-dimensional volume data of the fetal rib are identified from the three-dimensional volume data of the fetal rib structure.

In the second identification method, based on the first rib detection model, different fetal ribs or fetal spines are used as different identification objects, and the three-dimensional volume of the fetal spine is identified from the three-dimensional volume data of the first tissue to be measured. data and the three-dimensional volume data of the fetal rib.

S203. Render the 3D volume data of the fetal spine and the 3D volume data of the fetal ribs to obtain a 3D ultrasound image of the fetal rib structure;

After identifying the three-dimensional volume data of the fetal ribs and the three-dimensional volume data of the spine included in the first tissue to be measured in S202, three-dimensional rendering is performed on the three-dimensional volume data of the fetal spine and the three-dimensional volume data of the fetal ribs, Obtain a three-dimensional ultrasound image of the fetal rib structure.

Here, when performing three-dimensional rendering on the three-dimensional volume data of the fetal spine and the three-dimensional volume data of the fetal ribs, the three-dimensional volume data of the first tissue to be measured except the three-dimensional volume data of the fetal spine and the three-dimensional volume data of the fetal ribs All the data is cleared, and the three-dimensional volume data of the fetal spine and the three-dimensional volume data of the fetal ribs are rendered in stereoscopic light to form a three-dimensional ultrasound image of the fetal rib structure.

It should be noted that there may be multiple ways of rendering, and this embodiment of the present application does not impose any limitation on the way of rendering.

S204. Mark the fetal ribs in the three-dimensional ultrasound image of the fetal rib structure and/or determine the number of fetal ribs according to the identified three-dimensional volume data of the fetal ribs;

According to the position of the three-dimensional data of the fetal rib identified in S202 in the three-dimensional volume data of the first tissue to be measured, determine the position of the three-dimensional data of the fetal rib in the three-dimensional ultrasound image of the fetal rib structure in S203.

Here, the fetal ribs in the fetal rib structure can be marked according to the determined position of the fetal ribs, to obtain a three-dimensional ultrasonic image of the marked fetal rib structure, or to identify the number of fetal ribs from the three-dimensional strip data of fetal ribs. When marking the fetal ribs, one or more fetal ribs can be marked: for example: T1-T12, where T1 to T12 respectively represent the 1st to 12th fetal ribs, while marking the fetal ribs, Determine the number of fetal ribs.

In practical applications, when marking the fetal ribs, it can be judged whether the fetal rib structure includes abnormal fetal ribs with abnormal shapes or lesions according to the rib characteristics such as the shape characteristics and grayscale characteristics of the fetal ribs. If it is judged that the fetal rib structure includes abnormal fetal ribs, the abnormal fetal ribs can be marked with an abnormality mark, and the abnormal information of the abnormal fetal ribs relative to the normal fetal ribs can be determined.

S205. Outputting the marked three-dimensional ultrasound image of the fetal rib structure and/or displaying the determined number of fetal ribs.

At this time, the content displayed on the display of the ultrasonic imaging device is: the three-dimensional ultrasonic image of the fetal rib structure marked with the fetal rib, the data of the fetal rib, or the marked three-dimensional image and the number of fetal ribs are displayed simultaneously, so that the user can intuitively View the fetal rib structure and/or data of the fetus more accurately, and distinguish the fetal ribs in the fetal rib structure.

In an embodiment, the method further includes: straightening the 3D volume data of the fetal ribs to obtain the 3D volume data of the straightened ribs; straightening the 3D volume data of the fetal spine to obtain the 3D volume data of the straightened spine; The straightened ribs in the three-dimensional volume data of straight ribs and the straightened fetal spine in the three-dimensional volume data of straightened spine are fitted by plane fitting to obtain the first plane; The image data on the first plane is obtained to obtain a coronal image of the rib structure of the fetus.

After the three-dimensional volume data of the fetal ribs are determined, the fetal ribs are straightened according to the three-dimensional volume data of the fetal ribs, so that the curved fetal ribs are straightened to obtain straightened ribs. After the three-dimensional volume data of the fetal spine is determined, the fetal ribs are straightened according to the three-dimensional volume data of the fetal spine, so that the curved fetal spine is straightened to obtain a straightened fetal spine. At this time, the first plane is determined by solving mathematical equations or plane fitting methods such as the least square method or Hough transformation, and the first plane is the coronal plane where the longitudinal axis of the straightened ribs and straightened fetal spine is located. The three-dimensional volume data of each fetal rib on the first plane and the three-dimensional volume data of the fetal spine are obtained to obtain the coronal plane of the fetal rib structure.

In practical applications, after the first plane is determined, the image data of a pixel point on the first plane is obtained according to the three-dimensional volume data of each straightened rib and the three-dimensional volume data of the straightened fetal spine, so as to obtain the straightened fetal rib structure coronal images of .

In one embodiment, the method further includes: determining the target fetal rib from the ribs of the fetal rib structure; straightening the 3D volume data of the target fetal rib to obtain the 3D volume data of the straightened target fetal rib; The three-dimensional volume data of the spine is straightened to obtain the three-dimensional volume data of the straightened spine; the plane that is on the same plane as the straightened target rib and perpendicular to the straightened fetal spine is used as the second plane; according to the three-dimensional volume data of the straightened target fetal rib and the straightened The image data on the second plane is obtained from the three-dimensional volume data of the straight spine, and the cross-section of the rib of the target fetus is obtained.

Here, the target fetal ribs may include all the fetal ribs in the fetal rib structure, and may also be a part of the fetal ribs in the fetal rib structure. The target fetal rib can be selected by the user or automatically determined by the system. For example: display the volume data of the fetal rib structure on the display, determine the target fetal rib according to the rib identifier manually input by the user, or determine the target fetal rib based on the rib selection by the user on the fetal rib structure.

After the target fetal rib is determined, the target fetal rib is straightened according to the three-dimensional volume data of the target fetal rib, so that the curved target fetal rib is straightened to obtain the straightened target rib. After the three-dimensional volume data of the fetal spine is determined, the fetal spine is straightened according to the three-dimensional volume data of the fetal spine, so that the curved fetal spine is straightened to obtain a straightened fetal spine. At this time, a second plane where the target fetal rib is located is determined by two straight lines defining a plane, wherein the second plane is perpendicular to the longitudinal axis of the straightened fetal spine and is coplanar with the longitudinal axis of the straightened target rib. After the second plane is determined, the three-dimensional volume data of the target fetal ribs and the three-dimensional volume data of the fetal spine on the second plane are acquired to obtain the cross-section of the fetal rib structure.

Herein, the "longitudinal axis" of fetal ribs, fetal spine, fetal lumbar spine or other tissues may refer to the axis along the length direction of the tissue, for example, it may be the central axis or other axes along the length direction.

In practical applications, after the second plane is determined, the 3D volume data of the target straightened ribs and the 3D volume data of the straightened fetal spine on the second plane are acquired to obtain the cross-section of the straightened fetal rib structure.

It should be noted that the number of obtained cross-sections corresponds to the number of fetal ribs in the target fetal ribs, and different fetal ribs correspond to different cross-sections.

In one embodiment, straightening the straightened object includes: determining the longitudinal axis of the straightened object; performing equal interval sampling on the straightened object according to the longitudinal axis of the straightened object to obtain an equally spaced slice sequence ; the straightened objects include different fetal ribs or the fetal spine; based on the equal intervals, the sequence of equally spaced slices is reconstructed along a straight line.

Straightening The straightening of the object is divided into two parts: the extraction of the longitudinal axis and the reconstruction of straightening. The extraction method of the longitudinal axis can be: a tracking-based longitudinal axis extraction algorithm, a model-based multi-scale longitudinal axis extraction algorithm, a morphology-based longitudinal axis extraction method, a region growth-based centerline extraction method, a method based on three-dimensional geometric moments And the method of using machine learning to locate the centerline, etc.

For example, the tracking-based longitudinal axis extraction algorithm is a semi-automatic algorithm. Based on the initial key point and end point provided by the user, the tangent plane perpendicular to the tracking direction is generated during the process of tracking and straightening the object, and the maximum likelihood method and the centroid The method accurately calculates the center point of the straightened object in the cut plane, and after the tracking process is completed, the longitudinal axis of the straightened object is obtained by interpolating the center point sequence.

For another example, when the model-based multi-scale longitudinal axis extraction algorithm extracts the longitudinal axis of the straightening object, the local straightening object is approximated as a tubular structure, and the center of gravity of the tubular structure obtained by calculating the geometric moments is used as the center of the local straightening object. By analyzing the eigenvalues of the Hessian matrix corresponding to the multi-scale Gaussian filtered lower body data, the local straightening object is enhanced, and the longitudinal axis direction of the straightening object is estimated according to the eigenvector corresponding to the minimum eigenvalue of the Hessian matrix eigenvalue.

After the longitudinal axis of the straightened object is determined, the straightened object is straightened and reconstructed. First, the longitudinal axis of the rib is sampled at equal intervals to obtain the center points of equal intervals. Straightens the sequence of slices in the direction of the object's longitudinal axis. The obtained slice sequences are stacked together, and three-dimensional reconstruction is performed on the equally spaced slice sequences to obtain the straightened object after straightening.

In one embodiment, the method further includes: separating the 3D volume data of the fetal ribs and the 3D volume data of the fetal spine from the 3D volume data of the first tissue to be measured; Binarize the three-dimensional volume data of the ribs and the three-dimensional volume data of the fetal spine to obtain reconstructed three-dimensional volume data; render the reconstructed three-dimensional volume data to obtain the three-dimensional skeleton of the fetal rib structure.

After identifying the 3D volume data of the fetal ribs and the 3D volume data of the fetal spine, binarize the 3D volume data of the fetal ribs and the 3D volume data of the spine to obtain reconstructed 3D volume data, for example: if it is the fetal rib structure The local grayscale is set to 1, and the volume data other than the fetal rib structure is set to 0. The reconstructed 3D volume data is rendered by means of volume rendering or surface rendering to obtain a 3D skeleton including fetal ribs and spine.

In practical applications, when identifying the spine, only the three-dimensional volume data of the vertebra connected to the ribs in the spine can be identified. The rib structure including each fetal rib and vertebrae is thereby output.

In the ultrasonic imaging method provided in the embodiment of the present application, after obtaining the three-dimensional volume data of the fetus, the stereoscopic virtual display (Virtual Reality, VR) image of the fetal rib structure is automatically imaged to obtain a three-dimensional ultrasonic image, and the three-dimensional ultrasonic image is obtained. The fetal ribs are marked in the fetal rib structure, which greatly simplifies the workflow of fetal rib inspection. Further, the three-dimensional extraction of the rib skeleton is performed, and the straightened cross-section and coronal plane of all ribs or specified ribs are automatically imaged. It frees doctors from tedious and complicated manual operations, reduces dependence on doctors' skills, and improves inspection efficiency.

An embodiment of the present application provides an ultrasonic imaging method, which is applied to the ultrasonic imaging device shown in Figure 1, as shown in Figure 3, the method includes:

S301. Acquire three-dimensional volume data of the first tissue to be measured;

The first tissue to be tested includes: fetal rib structure and tissues other than fetal rib structure, such as amniotic fluid region, placenta, uterine wall and other tissues.

The doctor can scan the pregnant woman through the probe to obtain the three-dimensional volume data of the first tissue to be measured.

Here, the tissues other than the fetal rib structure belong to the hypoechoic region, and are low grayscale data in the three-dimensional volume data, and the fetal rib structure belongs to the hyperechoic region, and are high grayscale data in the three-dimensional volume data.

S302. Based on the first rib detection model, take different fetal ribs or fetal spines as different identification objects, and identify the three-dimensional volume data of the fetal spine and the three-dimensional volume data of the first tissue to be measured from the three-dimensional volume data of the first tissue to be measured. Three-dimensional volume data of fetal ribs;

The algorithm adopted by the first rib detection model can be a machine learning method. The first rib detection model uses the three-dimensional data of each fetal rib and the three-dimensional data of the spine as training samples, and learns the training samples through machine learning methods to pass the training. The sample trains the first rib detection model. The trained first rib detection model learns the volume data features of the fetal ribs and the spine. The volume data features can include: Principal Component Analysis (PCA) features, Linear Discriminant Analysis (Linear Discriminant Analysis, , LDA) features, Harr features, texture features and other features.

When the first rib detection model receives the three-dimensional volume data of the first tissue to be measured, according to the learned volume data features of the fetal ribs and the volume data features of the spine, identify the included fetal ribs from the received three-dimensional volume data 3D volume data and 3D volume data of the spine.

In the training samples, the fetal rib or spine is used as the target, and the target is calibrated, and the category of each calibrated target is indicated. Among them, the calibration can be performed by means of a Region Of Interest (ROI) box containing the target, or by a mask (Mask) that accurately segments the target.

The algorithm adopted by the first rib detection model can be an image segmentation algorithm, which performs binary segmentation on the 3D volume data input to the first rib detection model, and performs operations such as morphology, contour extraction, and domain connection to obtain multiple candidate regions. According to the volume data characteristics of each candidate area, the probability that each candidate area is the fetal rib or spine is judged, and the candidate area with the highest probability is selected as the area corresponding to the rib or spine, and the three-dimensional volume data of the selected area is the fetal rib or spine. 3D volume data.

In practical applications, the image segmentation algorithm adopted by the first rib detection model can also be: Level Set, Graph Cut, Snake, Random walker, active contour model algorithm, active shape model One or more of algorithms, active appearance model algorithms, and image segmentation algorithms in deep learning such as fully convolutional networks (FullyConvolutional Networks, FCN) and UNet.

The algorithm adopted by the first rib detection model may also be a template matching algorithm, which establishes a template for the three-dimensional volume data of the fetal ribs or the three-dimensional volume data of the spine. The first rib detection model performs binary segmentation on the input 3D volume data, and performs operations such as morphology, contour extraction, and Unicom domain to obtain multiple candidate areas, and traverses all candidate areas in the volume data according to the established template, and Determine the similarity between all candidate regions and the template, select the candidate region with the highest similarity as the target region corresponding to the fetal rib or spine, and the 3D volume data of the target region is the 3D volume data of the fetal rib or the 3D volume data of the spine.

It should be noted that the first rib model can directly count the number of fetal ribs included in the first tissue to be measured when identifying different fetal ribs or fetal spines.

S303. Render the 3D volume data of the fetal spine and the 3D volume data of the fetal ribs to obtain a 3D ultrasound image of the fetal rib structure;

After identifying the three-dimensional volume data of the fetal ribs and the three-dimensional volume data of the spine included in the first tissue to be measured in S302, three-dimensional rendering is performed on the three-dimensional volume data of the fetal spine and the three-dimensional volume data of the fetal ribs, Obtain a three-dimensional ultrasound image of the fetal rib structure.

Here, when performing three-dimensional rendering on the three-dimensional volume data of the fetal spine and the three-dimensional volume data of the fetal ribs, the three-dimensional volume data of the first tissue to be measured except the three-dimensional volume data of the fetal spine and the three-dimensional volume data of the fetal ribs All the data of the fetal spine are cleared, and the three-dimensional volume data of the fetal spine and the three-dimensional volume data of the fetal ribs are rendered in stereo light to form a three-dimensional ultrasound image of the fetal rib structure.

It should be noted that there may be multiple ways of rendering, and this embodiment of the present application does not impose any limitation on the way of rendering.

S304. Mark the fetal ribs in the three-dimensional ultrasound image of the fetal rib structure;

According to the position in the three-dimensional data of the fetal rib identified in S302 in the three-dimensional volume data of the first tissue to be measured, determine the position of the three-dimensional data of the fetal rib in the three-dimensional ultrasound image of the fetal rib structure in S203, and according to the determined position The position of the fetal rib is marked on the fetal rib in the fetal rib structure, and a three-dimensional ultrasonic image of the marked fetal rib structure is obtained.

When marking the fetal ribs, one or more rib identifications of the fetal ribs can be marked: for example: T1-T12, where T1 to T12 represent the 1st to 12th fetal ribs respectively.

S305. Outputting the marked three-dimensional ultrasonic image of the rib structure of the fetus.

At this time, the content displayed on the display of the ultrasonic imaging device is: a three-dimensional ultrasonic image of the fetal rib structure marked with fetal ribs, so that the user can visually view the fetal rib structure of the fetus and distinguish the fetal rib structure from the fetal rib structure .

In the ultrasonic imaging method provided in the embodiment of the present application, after obtaining the three-dimensional volume data of the fetus, the fetal ribs and spine are taken as different identification objects, the fetal ribs and the fetal spine are identified, and the first tissue to be measured is included in the statistics automatically. The number of fetal ribs can be automatically imaged on the three-dimensional VR map of the fetal rib structure, which greatly simplifies the workflow of fetal rib inspection, liberates doctors from tedious and complicated manual operations, reduces the dependence on doctors' skills, and reduces rate of misdiagnosis and missed diagnosis.

An embodiment of the present application provides an ultrasonic imaging method, which is applied to the ultrasonic imaging device shown in FIG. 1, as shown in FIG. 4, and the method includes:

S401. Obtain three-dimensional volume data of the first tissue to be measured;

The first tissue to be tested includes: fetal rib structure and tissues other than fetal rib structure, such as amniotic fluid region, placenta, uterine wall and other tissues.

The doctor can scan the pregnant woman through the probe to obtain the three-dimensional volume data of the first tissue to be measured.

Here, the tissues other than the fetal rib structure belong to the hypoechoic region and appear as low gray in the ultrasound image, and the fetal rib structure belongs to the hyperechoic region and appear as high gray in the ultrasound image.

S402. Identify the three-dimensional volume data of the fetal rib structure from the three-dimensional volume data of the first tissue to be measured;

When identifying the three-dimensional volume data of the fetal rib structure from the three-dimensional volume data of the first tissue to be measured, at least one of the following three structure identification methods may be included:

Structure identification method 1: displaying the 3D image corresponding to the 3D volume data of the first tissue to be measured; receiving a first input operation based on the 3D volume data corresponding to the 3D volume data of the first tissue to be measured; determining the first input operation Corresponding marker points: identifying the three-dimensional volume data of the fetal rib structure from the three-dimensional volume data of the first tissue to be measured according to the coordinates of the marker points.

After the three-dimensional volume data of the first tissue to be measured is acquired, a three-dimensional image corresponding to the three-dimensional volume data of the first tissue to be measured is displayed on the display. Based on tools such as a trackball and a touch screen, the user performs the first input by selecting marker points, drawing marker lines, and other methods in the fetal rib structure on the three-dimensional image corresponding to the three-dimensional volume data of the first tissue to be measured displayed on the display. operation to inform the ultrasound imaging equipment of the location of the fetal rib structures in space. Here, the marker line is composed of a plurality of consecutive marker points. After receiving the first input operation, the ultrasonic imaging device determines the position of the fetal rib structure in the three-dimensional volume data of the first tissue to be measured through the coordinates of the corresponding marker points or the marker points constituting the marker line of the first input operation.

For example: the user picks some points intermittently on the endpoints of each fetal rib or on the boundary of the fetal rib structure. The ultrasound imaging device takes the points selected by the user as marker points, and roughly draws the midline or the centerline of the fetal rib structure according to the coordinates of the marker points. The boundary line of the fetal rib structure is traced, and the position of the fetal rib structure is determined according to the determined midline or boundary line, so as to obtain the three-dimensional volume data of the fetal rib structure.

The second structure recognition method: displaying the 3D image corresponding to the 3D volume data of the first tissue to be measured; receiving a second input operation based on the 3D image corresponding to the 3D volume data of the first tissue to be measured; determining the second input Operating the corresponding first seed area; the first seed area is located in the three-dimensional image area corresponding to the fetal rib structure; determining the first pixel feature of the three-dimensional volume data of the first seed area; according to the first pixel feature in The three-dimensional volume data of the fetal rib structure is identified in the three-dimensional volume data of the first tissue to be measured.

Here, after the three-dimensional volume data of the first tissue to be measured is acquired, the three-dimensional image corresponding to the three-dimensional volume data of the first tissue to be measured is displayed on the display. The user performs a second input operation on the three-dimensional image corresponding to the three-dimensional volume data of the first tissue to be measured displayed on the display by using tools such as a trackball and a touch screen. After receiving the second input operation, the ultrasonic imaging device determines that the area corresponding to the second input operation is the first seed area, uses the three-dimensional volume data of the first seed area as the prior data, and obtains the edge gradient and grayscale of the prior data The isopixel feature is used as the first pixel feature, and the three-dimensional volume data of the fetal rib structure is identified from the three-dimensional volume data of the first tissue to be measured according to the first pixel feature of the prior data.

The method for identifying the three-dimensional volume data of the fetal rib structure from the three-dimensional volume data of the first tissue to be measured according to the pixel features of the prior data may include: one of template matching, image feature extraction, edge extraction, and morphological operations. One or more image processing methods may also include: one or more image segmentation in graphcut algorithm, grabcut algorithm, level set method, active contour model algorithm, active shape model algorithm, seed region growth method, region segmentation and merging method The method may also include: one or more machine learning methods in deep learning method, support vector machine, adaboost, random forest algorithm.

For example, when the three-dimensional volume data of the fetal rib structure is identified from the three-dimensional volume data of the first tissue to be measured by template matching, the prior data is used as a template to calculate pixel features such as edge gradients and gray levels of the template, and the template The three-dimensional volume data of the first tissue to be measured is traversed to find the optimal solution with the smallest difference from the pixel features of the template, so as to realize the identification of the fetal rib structure.

Another example: when the three-dimensional volume data of the fetal rib structure is identified from the three-dimensional volume data of the first tissue to be measured by the seed area growing method, the seed area is determined in the area of the fetal rib structure, and the pixels of the seed area are used as the seed pixels , and then according to the first pixel feature of the seed pixel, from the 3D volume data of the first tissue to be measured, the pixels conforming to the first pixel feature are merged into the seed region, and the newly added pixel is used as a new seed pixel to continue merging until Until no new pixel meeting the conditions can be found, the three-dimensional volume data of the fetal rib structure is finally recognized.

Here, certain prior data is pre-acquired in the fetal rib structure data as known information through a certain workflow of user interaction, and the difficulty of identifying the fetal rib structure is reduced through the first pixel feature of the known information.

Structure identification method 3: Determine at least two first candidate areas from the three-dimensional volume data of the first tissue to be measured, and acquire the volume data features of the three-dimensional volume data of each first candidate area; The volume data feature of the candidate area is to determine the first matching degree of each first candidate area and the fetal rib structure; determine the first candidate area with the highest first matching degree as the target area corresponding to the fetal rib structure; The three-dimensional volume data of the target area corresponding to the fetal rib structure is used as the three-dimensional volume data of the fetal rib structure.

Here, the three-dimensional volume data of the first tissue to be measured may be received through the rib structure detection model, and the rib structure detection model determines at least two first candidate regions from the three-dimensional volume data of the first tissue to be measured, and obtains each first The volume data feature of the three-dimensional volume data of the candidate area; the rib structure detection model determines the first matching degree between each first candidate area and the rib structure of the fetus according to the volume data feature of each first candidate area; the rib structure The detection model determines that the first candidate area with the highest matching degree is the target area corresponding to the fetal rib structure; the rib structure detection model uses the three-dimensional volume data of the target area corresponding to the fetal rib structure as the three-dimensional data of the fetal rib structure volume data.

The algorithm adopted by the rib structure detection model can be a machine learning method. The rib structure detection model uses the three-dimensional data of the rib structure of each fetus as a training sample, and uses the machine learning method to learn the training samples, so as to perform the rib structure detection model through the training samples. train. The trained rib structure detection model learns the volume data features of the fetal rib structure, wherein the volume data features may include: PCA features, LDA features, Harr features, texture features and other features.

When the rib structure detection model receives the 3D volume data of the first tissue to be measured, it recognizes the included 3D volume data of the fetal rib structure from the received 3D volume data according to the learned volume data features of the fetal rib structure.

In the training samples, the fetal rib structure is used as the target, the target is calibrated, and the category of each calibrated target is pointed out. Among them, the calibration can be performed by means of a Region Of Interest (ROI) box containing the target, or by a mask (Mask) that accurately segments the target.

The algorithm used by the rib structure detection model can be an image segmentation algorithm, which performs binary segmentation on the 3D volume data input to the rib structure detection model, and performs operations such as morphology, contour extraction, and connected domains to obtain multiple first candidate regions. According to the volume data characteristics of each first candidate region, the probability that each first candidate region is the fetal rib structure is judged, and the first candidate region with the highest probability is selected as the target region corresponding to the fetal rib structure, and the three-dimensional The volume data is three-dimensional volume data of fetal rib structure.

In practical applications, the image segmentation algorithm used by the rib structure detection model can also be: Level Set, Graph Cut, Snake, Random walker, Active Contour Model Algorithm, Active Shape Model Algorithm , one or more of active appearance model algorithms, and image segmentation algorithms in deep learning such as FCN and UNet.

The algorithm adopted by the rib structure detection model may also be a template matching algorithm to establish a template of the three-dimensional volume data of the fetal rib structure. The rib structure detection model performs binary segmentation on the input 3D volume data, and performs operations such as morphology, contour extraction, and Unicom domain to obtain multiple first candidate areas, and traverses all the first candidates in the volume data according to the established template region, and determine the similarity between all the first candidate regions and the template, select the first candidate region with the highest similarity as the target region corresponding to the fetal rib structure, and the three-dimensional volume data of the target region is the three-dimensional volume data of the fetal rib structure.

It should be noted that, in the embodiment of the present application, the specific identification manner of identifying the three-dimensional volume data of the fetal rib structure from the three-dimensional volume data of the first tissue to be measured is not limited in any way.

S403. Segment the three-dimensional volume data of the fetal rib structure from the three-dimensional volume data of the first tissue to be measured;

After identifying the three-dimensional volume data of the fetal rib structure, according to the position of the three-dimensional volume data of the fetal rib structure in the three-dimensional volume data of the first tissue to be measured, the three-dimensional volume data of the fetal rib structure are transferred from the three-dimensional volume data of the first tissue to be measured to separated from the data.

In practical application, S403 and S404 can be realized simultaneously, and at the same time of identifying the three-dimensional volume data of the fetal rib structure, the three-dimensional volume data of the fetal rib structure is segmented from the three-dimensional volume data of the first tissue to be measured.

S404. Identify the three-dimensional volume data of the fetal spine and the three-dimensional volume data of the fetal ribs from the three-dimensional volume data of the fetal rib structure;

In S403, after separating the 3D volume data of the fetal rib structure from the 3D volume data of the first tissue to be measured, when identifying the 3D volume data of the fetal ribs and the 3D volume data of the spine from the 3D volume data of the fetal rib structure , which can include at least one of the following two rib identification methods:

Rib recognition method 1. Based on the shape characteristics of the fetal spine and the shape characteristics of the fetal ribs, locate the included spine and fetal ribs from the three-dimensional volume data of the fetal rib structure; display the corresponding A three-dimensional image; receiving a third input operation based on the three-dimensional image corresponding to the three-dimensional volume data of the fetal rib structure; according to the third input operation, determining a reference fetal rib and a rib identification corresponding to the reference fetal rib; the reference fetal The rib is at least one fetal rib in the fetal rib structure; based on the reference fetal rib and the rib identification corresponding to the reference fetal rib, identify the three-dimensional volume data of the fetal rib.

After the three-dimensional volume data of the fetal rib structure is determined, a three-dimensional image corresponding to the three-dimensional volume data of the fetal rib structure is displayed on the display. Based on tools such as a trackball and a touch screen, the user performs a third input operation on the three-dimensional image corresponding to the three-dimensional volume data of the rib structure displayed on the display to identify some fetal ribs in the fetal rib structure and the identified fetal ribs. Rib identification for rib structure.

After receiving the third input operation, the ultrasonic imaging device determines the fetal rib identified by the user through the third input operation, takes the fetal rib identified by the user as the reference fetal rib, and uses the rib identification input by the user as the rib identification of the reference fetal rib, according to With reference to the fetal ribs, each fetal rib in the fetal rib structure and the rib identification corresponding to each fetal rib are determined. Here, the reference fetal ribs may include one or more fetal ribs.

When the ultrasonic imaging equipment determines each fetal rib in the fetal rib structure and the rib identification corresponding to each fetal rib according to the reference fetal rib, it can use the gray histogram projection method, contour extraction algorithm, edge extraction algorithm, Unicom domain method, group One or more methods of block detection algorithm, template matching algorithm, image feature extraction algorithm, and morphological operation algorithm determine the fetal ribs included in the fetal rib structure and the number of included fetal ribs.

For example: the user selects T3 and T9 ribs on the three-dimensional image corresponding to the fetal rib structure through tools such as trackball and touch screen, and clicks a point on T3 and T9 ribs or draws the centerlines of T3 and T9 ribs respectively. Ribs are marked as reference fetal ribs. Project the coronal plane of the three-dimensional volume data of the fetal rib structure through the histogram grayscale projection method along the Y axis to obtain a pixel statistical map, and calculate the number of pixel peaks and the position of the peak in the pixel statistical map, so as to obtain the fetal rib structure. quantity and location. The fetal rib above the T2 rib is determined as T1, the fetal rib below the T2 rib is determined as T3, and similarly, T4-T12 is determined. Therefore, the fetal ribs in the fetal rib structure and the rib identification of each fetal rib are determined according to the sequence order and spatial position relationship of the reference fetal ribs and the identified fetal ribs.

Rib identification method 2. Based on the second rib detection model, different fetal ribs or fetal spines are used as different identification objects, and the three-dimensional volume data and the fetal spine are identified from the three-dimensional volume data of the fetal rib structure. Describe the three-dimensional volume data of fetal ribs.

The algorithm adopted by the second rib detection model can be a machine learning method. The second rib detection model uses the three-dimensional volume data of each fetal rib and the three-dimensional volume data of the spine as training samples, and learns the training samples through machine learning methods to pass the training. The sample trains the second rib detection model. The trained second rib detection model learns volume data features of fetal ribs and spine, wherein the volume data features may include: PCA features, LDA features, Harr features, texture features and other features.

When the second rib detection model receives the three-dimensional volume data of the fetal rib structure, it recognizes the included three-dimensional volume of the fetal ribs from the received three-dimensional volume data according to the learned volume data features of the fetal ribs and the volume data features of the spine data and the 3D volumetric data of the spine.

In the training samples, the fetal rib or spine is used as the target, and the target is calibrated, and the category of each calibrated target is indicated. Wherein, the calibration can be performed by means of an ROI box containing the target, or by a mask (Mask) which accurately segments the target.

The algorithm adopted by the second rib detection model can be an image segmentation algorithm, which performs binary segmentation on the 3D volume data input into the second rib detection model, and performs operations such as morphology, contour extraction, and Unicom domain to obtain multiple candidate regions. According to the volume data characteristics of each candidate area, the probability that each candidate area is the fetal rib or spine is judged, and the candidate area with the highest probability is selected as the area corresponding to the rib or spine, and the three-dimensional volume data of the selected area is the fetal rib or spine. 3D volume data.

In practical applications, the image segmentation algorithm adopted by the second rib detection model can also be: Level Set, Graph Cut, Snake, Random walker, active contour model algorithm, active shape model One or more of algorithms, active appearance model algorithms, and image segmentation algorithms in deep learning such as fully convolutional networks (FullyConvolutional Networks, FCN) and UNet.

The algorithm adopted by the second rib detection model may also be a template matching algorithm, which establishes a template for the three-dimensional volume data of the fetal ribs or the three-dimensional volume data of the spine. The second rib detection model performs binary segmentation on the input 3D volume data, and performs operations such as morphology, contour extraction, and Unicom domain to obtain multiple candidate areas, and traverses all candidate areas in the volume data according to the established template, and Determine the similarity between all candidate regions and the template, select the candidate region with the highest similarity as the target region corresponding to the fetal rib or spine, and the 3D volume data of the target region is the 3D volume data of the fetal rib or the 3D volume data of the spine.

The algorithm adopted by the second rib detection model can also be: one or more methods such as image edge extraction, histogram image gray projection statistics, image contour extraction, morphological processing, threshold segmentation, clump detection, etc. to directly calculate the fetal ribs The number of fetal ribs in the volumetric data of the structure.

It should be noted that, when the second rib detection model identifies different fetal ribs or fetal spines, it can directly count the number of fetal ribs included in the fetal rib structure.

S405. Render the 3D volume data of the fetal spine and the 3D volume data of the fetal ribs to obtain a 3D ultrasound image of the fetal rib structure;

After identifying the 3D volume data of the fetal ribs and the 3D volume data of the spine included in the first tissue to be measured in S404, perform 3D rendering on the 3D volume data of the fetal spine and the 3D volume data of the fetal ribs, Obtain a three-dimensional ultrasound image of the fetal rib structure.

Here, when performing three-dimensional rendering on the three-dimensional volume data of the fetal spine and the three-dimensional volume data of the fetal ribs, the three-dimensional volume data of the first tissue to be measured except the three-dimensional volume data of the fetal spine and the three-dimensional volume data of the fetal ribs All the data of the fetal spine are cleared, and the three-dimensional volume data of the fetal spine and the three-dimensional volume data of the fetal ribs are rendered in stereo light to form a three-dimensional ultrasound image of the fetal rib structure.

It should be noted that there may be multiple ways of rendering, and this embodiment of the present application does not impose any limitation on the way of rendering.

S406. Mark the fetal ribs in the three-dimensional ultrasound image of the fetal rib structure;

According to the position in the three-dimensional data of the fetal rib identified in S404 in the three-dimensional volume data of the first tissue to be measured, determine the position of the three-dimensional data of the fetal rib in the three-dimensional ultrasonic image of the fetal rib structure in S203, and according to the determined position The position of the fetal rib is marked on the fetal rib in the fetal rib structure, and a three-dimensional ultrasonic image of the marked fetal rib structure is obtained.

When marking the fetal ribs, one or more rib identifications of the fetal ribs can be marked: for example: T1-T12, where T1 to T12 represent the 1st to 12th fetal ribs respectively.

S407 , outputting the marked three-dimensional ultrasonic image of the fetal rib structure.

At this time, the content displayed on the display of the ultrasonic imaging device is: a three-dimensional ultrasonic image of the fetal rib structure marked with fetal ribs, so that the user can visually view the fetal rib structure of the fetus and distinguish the fetal rib structure from the fetal rib structure .

The ultrasonic imaging method provided in the embodiment of the present application automatically or semi-automatically or semi-automatically recognizes, locates, and segments the three-dimensional volume data of the fetal rib structure after obtaining the three-dimensional volume data of the fetus, and performs automatic imaging on the three-dimensional VR map of the fetal rib structure, and Automatically or semi-automatically count the number of fetal ribs, and mark the fetal ribs in the fetal rib structure on the three-dimensional VR map. Further, the three-dimensional extraction of the rib skeleton is performed, and the straightened cross-section and coronal plane of all ribs or specified ribs are automatically imaged. This greatly simplifies the workflow of fetal rib inspection, liberates doctors from cumbersome and complicated manual operations, reduces the dependence on doctors' skills, and improves inspection efficiency; moreover, the stability and imaging quality of rib count results are compared with manual Are in a better state, reducing the rate of misdiagnosis and missed diagnosis.

An embodiment of the present application provides an ultrasonic imaging method, which is applied to the ultrasonic imaging device shown in FIG. 1, as shown in FIG. 5, and the method includes:

S501. Obtain three-dimensional volume data of the second tissue to be measured;

The second tissue to be tested includes: fetal spine and tissues other than fetal spine, such as amniotic fluid region, placenta, uterine wall and other tissues.

The doctor can scan the pregnant woman through the probe to obtain the three-dimensional volume data of the second tissue to be measured.

Here, the fetal rib structures belong to the hyperechoic region and appear as high gray levels in the ultrasound image.

S502. Identify the three-dimensional volume data of the conus medullaris and the three-dimensional volume data of the lumbar spine from the three-dimensional volume data of the second tissue to be measured;

After the three-dimensional volume data of the second tissue to be measured is obtained based on S201, the three-dimensional volume data of the conus medullaris and the three-dimensional volume data of the lumbar spine are identified from the acquired three-dimensional volume data.

The conus medullaris can be identified from the three-dimensional volume data of the second tissue to be measured by using the conus medullaris detection model.

In an embodiment, the identifying the three-dimensional volume data of the conus medullaris from the three-dimensional volume data of the second tissue to be measured includes: determining at least two first Two candidate areas, obtaining volume data features of the three-dimensional volume data of each second candidate area; determining a second match between each second candidate area and the fetal spine according to the volume data features of each second candidate area degree; determine the second candidate area with the highest second matching degree as the target area corresponding to the conus medullaris; use the three-dimensional volume data of the target area corresponding to the conus medullaris as the three-dimensional volume data of the conus medullaris.

Here, the conus medullaris detection model determines at least two second candidate areas from the three-dimensional volume data of the second tissue to be measured, and obtains the volume data features of the three-dimensional volume data of each second candidate area; Describe the volume data characteristics of each second candidate region, and determine the second matching degree between each second candidate region and the fetal spine; the conus medullaris detection model determines that the second candidate region with the highest second matching degree is the conus medullaris Corresponding target area; the conus medullaris detection model takes the 3D volume data of the target area corresponding to the conus medullaris as the 3D volume data of the conus medullaris.

The algorithm adopted by the conus medullaris detection model can be a machine learning method. The conus medullaris detection model uses the three-dimensional volume data of the conus medullaris as a training sample, and the training samples are learned by the machine learning method, so as to train the conus medullaris detection model through the training samples. The trained conus medullaris detection model learns the volume data features of the conus medullaris, wherein the volume data features may include: PCA features, LDA features, Harr features, texture features and other features.

When the conus medullaris detection model receives the 3D volume data of the second tissue to be measured, it identifies the included 3D volume data of the conus medullaris from the received 3D volume data according to the learned volume data features of the conus medullaris.

In the training samples, the conus medullaris is used as the target, and the target is calibrated. Wherein, the calibration can be performed by means of an ROI box containing the target, or by a mask (Mask) which accurately segments the target.

The algorithm adopted by the conus medullaris detection model can be an image segmentation algorithm, which performs binary segmentation on the three-dimensional volume data input into the conus medullaris detection model, and performs operations such as morphology, contour extraction, and connected domains to obtain multiple second candidate regions. According to the volume data characteristics of each second candidate area, the probability that each second candidate area is the conus medullaris is judged, and the first candidate area with the highest probability is selected as the target area corresponding to the conus medullaris, and the three-dimensional volume data of the selected target area is the three-dimensional volume data of the conus medullaris.

In practical applications, the image segmentation algorithm adopted by the conus medullaris detection model can also be: Level Set, Graph Cut, Snake, Random walker, Active Contour Model Algorithm, Active Shape Model Algorithm , one or more of active appearance model algorithms, and image segmentation algorithms in deep learning such as FCN and UNet.

The algorithm adopted by the conus medullaris detection model may also be a template matching algorithm, which establishes a template of three-dimensional volume data of the conus medullaris. The conus medullaris detection model performs binary segmentation on the input 3D volume data, and performs operations such as morphology, contour extraction, and connected domains to obtain multiple second candidate areas, and traverses all the second candidates in the volume data according to the established template area, and determine the similarity between all the second candidate areas and the template, select the second candidate area with the highest similarity as the target area corresponding to the conus medullaris, and the three-dimensional volume data of the target area is the three-dimensional volume data of the fetal conus medullaris.

It should be noted that, in the embodiment of the present application, the specific identification manner of identifying the three-dimensional volume data of the conus medullaris from the three-dimensional volume data of the second tissue to be measured is not limited in any way.

Identify the three-dimensional volume data of the lumbar spine, including at least one of the following two identification methods:

Identification method 1: Firstly, the whole fetal spine is used as the identification object to identify the 3D volume data of the fetal spine from the 3D volume data of the second tissue to be measured, and then the 3D volume data of the lumbar spine are identified from the 3D volume data of the fetal spine.

The second identification method is to directly use the lumbar spine as the identification object to identify the lumbar spine from the three-dimensional volume data of the second tissue to be measured.

In identification mode 1, identify the three-dimensional volume data of the fetal spine from the three-dimensional volume data of the second tissue to be measured; Segmentation; identifying the three-dimensional volume data of the lumbar spine from the three-dimensional volume data of the fetal spine.

In the second identification method, based on the first lumbar detection model, the lumbar spine is used as an identification object, and the three-dimensional volume data of the lumbar spine are identified from the three-dimensional volume data of the second tissue to be measured.

In an embodiment, the identifying the 3D volume data of the lumbar spine from the 3D volume data of the second tissue to be measured includes: determining at least two third volume data from the 3D volume data of the second tissue to be measured Candidate areas, obtaining the volume data features of the three-dimensional volume data of each third candidate area; according to the volume data features of each third candidate area, determining the first position between each third candidate area and the lumbar vertebra in the fetal spine three matching degrees; determine the third candidate area with the highest third matching degree as the target area corresponding to the lumbar spine; use the 3D volume data of the target area corresponding to the lumbar spine as the 3D volume data of the lumbar spine.

Here, the first lumbar detection model determines at least two third candidate regions from the three-dimensional volume data of the second tissue to be measured, and obtains the volume data features of the three-dimensional volume data of each third candidate region; the first lumbar detection model According to the volume data characteristics of each third candidate area, determine the third matching degree between each third candidate area and the lumbar vertebra in the fetal spine; the first lumbar detection model determines the third candidate with the highest third matching degree The area is the target area corresponding to the lumbar spine; the first lumbar detection model uses the 3D volume data of the target area corresponding to the lumbar spine as the 3D volume data of the lumbar spine.

The algorithm adopted by the first lumbar vertebra detection model can be a machine learning method. The first lumbar vertebra detection model uses the three-dimensional data of the lumbar vertebra as a training sample, and learns the training samples through the machine learning method, so as to perform the first lumbar vertebra detection model through the training samples. train. The trained first lumbar detection model learns volume data features of the lumbar spine, wherein the volume data features may include: PCA features, LDA features, Harr features, texture features and other features.

When the first lumbar detection model receives the 3D volume data of the second tissue to be measured, it identifies the 3D volume data of the lumbar spine included in the received 3D volume data according to the learned features of the lumbar volume data.

S503. Render the 3D volume data of the conus medullaris and the 3D volume data of the lumbar spine to obtain a 3D ultrasound image of the vertebral structure;

After identifying the 3D volume data of the conus medullaris and the 3D volume data of the lumbar spine in the second tissue to be measured in S502, three-dimensional rendering is performed on the 3D volume data of the conus medullaris and the 3D volume data of the lumbar spine to obtain a 3D ultrasound image of the vertebral structure.

Here, when three-dimensional rendering is performed on the three-dimensional volume data of the conus medullaris and the three-dimensional volume data of the lumbar spine, all the data except the three-dimensional volume data of the conus medullaris and the three-dimensional volume data of the lumbar spine in the three-dimensional volume data of the second tissue to be measured are Empty, the three-dimensional volume data of the spinal cord and the three-dimensional volume data of the lumbar spine are rendered in stereo light perspective to form a three-dimensional ultrasound image of the vertebral structure.

It should be noted that there may be multiple ways of rendering, and this embodiment of the present application does not impose any limitation on the way of rendering.

S504. Mark the conus medullaris in the three-dimensional ultrasound image of the vertebral structure;

According to the position of the three-dimensional data of the conus medullaris identified in S502 in the three-dimensional volume data of the second tissue to be measured, determine the position of the three-dimensional data of the conus medullaris in the three-dimensional ultrasonic image of the fetal rib structure in S203, and according to the determined position of the spinal cord The position of the cone is used to mark the conus medullaris, and a three-dimensional ultrasound image of the marked vertebral structure is obtained.

Here, the displayed three-dimensional ultrasonic image of the vertebral structure may include only the conus medullaris and the lumbar vertebrae, or may include vertebrae such as vertebrae other than the lumbar vertebrae.

In practical application, the distance between the conus medullaris and the lumbar vertebra is detected, and the detected distance is displayed on the display screen. The detected distance is compared with the set distance range, and if the detected distance does not match the set distance range, a prompt message is issued to prompt the user that there is an abnormality in the current position of the conus medullaris.

S505. Output the marked three-dimensional ultrasonic image of the vertebral structure.

At this time, the content displayed on the monitor of the ultrasound imaging device is: a three-dimensional ultrasound image of the vertebral structure marked with the conus medullaris, enabling the user to visually view the position of the conus medullaris of the fetus and determine the distance between the conus medullaris and the lumbar spine .

In one embodiment, marking the conus medullaris in the three-dimensional ultrasound image of the vertebral structure comprises: determining the three-dimensional coordinates of the end of the conus medullaris; determining the longitudinal axis of the lumbar spine; The three-dimensional coordinates of the fetus and the longitudinal axis of the fetal spine determine a reference line or a third plane passing through the end of the conus medullaris and perpendicular to the longitudinal axis of the lumbar spine; mapping the reference line or the third plane on the 3D ultrasound image of vertebral structures.

According to the two-dimensional coordinates of the end of the conus medullaris on the sagittal plane and the specific position of the sagittal plane in the three-dimensional volume data, the three-dimensional coordinates of the end of the conus medullaris in the three-dimensional volume data are calculated. The longitudinal axis in the three-dimensional volume data is calculated as a plane or a straight line that passes through the end point of the conus medullaris and is perpendicular to the lumbar spine, and this plane or the straight line is called the third plane. By mapping the third plane to the stereoscopic VR map of the fetal spine, the relative position of the end of the conus medullaris to the fetal lumbar vertebrae can be intuitively expressed, and the calculated result is marked on the VR map. Wherein, the longitudinal axis of the lumbar spine in the three-dimensional volume data represents the posture of the fetal spine.

The ultrasound imaging method provided in the embodiment of the present application automatically images the stereoscopic VR image of the fetal spine after obtaining the three-dimensional volume data of the fetus, and marks the conus medullaris on the stereoscopic VR image, thereby greatly simplifying the examination of the conus medullaris workflow. Further, the specific position of the conus medullaris in the lumbar spine was marked on the stereoscopic VR map. It frees doctors from tedious and complicated manual operations, reduces dependence on doctors' skills, and improves inspection efficiency.

An embodiment of the present application provides an ultrasonic imaging method, which is applied to the ultrasonic imaging device shown in FIG. 1, as shown in FIG. 6, and the method includes:

S601. Acquire three-dimensional volume data of the second tissue to be measured;

The second tissue to be tested includes: fetal spine and tissues other than fetal spine, such as amniotic fluid region, placenta, uterine wall and other tissues.

The doctor can scan the pregnant woman through the probe to obtain the three-dimensional volume data of the second tissue to be measured.

Here, the fetal rib structures belong to the hyperechoic region and appear as high gray levels in the ultrasound image.

S602. Identify the three-dimensional volume data of the conus medullaris from the three-dimensional volume data of the second tissue to be measured;

After the three-dimensional volume data of the second tissue to be measured is obtained based on S601, the three-dimensional volume data of the conus medullaris is identified from the acquired three-dimensional volume data.

The conus medullaris can be identified from the three-dimensional volume data of the second tissue to be measured by using the conus medullaris detection model. The conus medullaris detection model determines at least two second candidate regions from the three-dimensional volume data of the second tissue to be measured, and obtains the volume data characteristics of the three-dimensional volume data of each second candidate region; the conus medullaris detection model Volume data characteristics of a second candidate area, determining the second matching degree between each second candidate area and the fetal spine; the conus medullaris detection model determines that the second candidate area with the highest second matching degree is corresponding to the conus medullaris Target area; the conus medullaris detection model takes the 3D volume data of the target area corresponding to the conus medullaris as the 3D volume data of the conus medullaris.

S603. Identify the three-dimensional volume data of the fetal spine from the three-dimensional volume data of the second tissue to be measured;

After obtaining the 3D volume data of the second tissue to be measured based on S601 , identify the 3D volume data of the fetal spine from the acquired 3D volume data.

When identifying the three-dimensional volume data of the fetal spine from the three-dimensional volume data of the second tissue to be measured, at least one of the following three spine identification methods may be included:

Spine identification method 1: displaying the 3D image corresponding to the 3D volume data of the second tissue to be measured; receiving a fourth input operation based on the 3D volume data corresponding to the 3D volume data of the second tissue to be measured; determining the fourth input operation Corresponding marker points: According to the coordinates of the marker points corresponding to the fourth input operation, identify the three-dimensional volume data of the spine from the three-dimensional volume data of the second tissue to be measured.

After acquiring the 3D volume data of the second tissue to be measured, a 3D image corresponding to the 3D volume data of the first tissue to be measured is displayed on the display. Based on tools such as a trackball and a touch screen, the user performs a fourth input operation in the fetal spine on the three-dimensional image corresponding to the three-dimensional volume data of the second tissue to be measured displayed on the display, by selecting marker points, drawing marker lines, etc. , to inform the ultrasound imaging equipment of the position of the fetal spine in space. Here, the marker line is composed of a plurality of consecutive marker points. After receiving the fourth input operation, the ultrasonic imaging device determines the position of the fetal spine in the three-dimensional volume data of the second tissue to be measured through the coordinates of the corresponding marker points or the marker points forming the marker line in the fourth input operation.

For example: the user picks some points intermittently on the endpoints of each vertebra of the fetal spine or on the boundary of the fetal spine, and the ultrasound imaging device takes the points selected by the user as marker points, and roughly draws the midline of the fetal spine according to the coordinates of the marker points Or trace the boundary line of the fetal spine, and determine the position of the fetal spine according to the determined midline or boundary line, so as to obtain the three-dimensional volume data of the fetal spine.

Spine identification method 2: displaying the 3D image corresponding to the 3D volume data of the second tissue to be measured; receiving a fifth input operation based on the 3D image corresponding to the 3D volume data of the second tissue to be measured; determining the fifth input Operating the corresponding second seed area; the second seed area is located in the three-dimensional image area corresponding to the spine; determining the second pixel feature of the three-dimensional volume data of the second seed area; according to the second pixel feature in the The three-dimensional volume data of the spine is identified from the three-dimensional volume data of the second tissue to be measured.

Here, after the three-dimensional volume data of the second tissue to be measured is acquired, the three-dimensional image corresponding to the three-dimensional volume data of the second tissue to be measured is displayed on the display. The user performs a fifth input operation on the 3D image corresponding to the 3D volume data of the second tissue to be measured displayed on the display by using tools such as a trackball and a touch screen. After receiving the fifth input operation, the ultrasonic imaging device determines that the area corresponding to the fifth input operation is the second seed area, uses the three-dimensional volume data of the second seed area as the prior data, and obtains the edge gradient and grayscale of the prior data The equal pixel feature is used as the second pixel feature, and the three-dimensional volume data of the fetal spine is identified from the three-dimensional volume data of the second tissue to be measured according to the second pixel feature of the prior data.

The method for identifying the three-dimensional volume data of the fetal spine from the three-dimensional volume data of the second tissue to be measured according to the pixel characteristics of the prior data may include: one of template matching, image feature extraction, edge extraction, and morphological operations. or multiple image processing methods, may also include: one or more image segmentation methods in graphcut algorithm, grabcut algorithm, level set method, active contour model algorithm, active shape model algorithm, seed region growth method, region segmentation and merging method , may also include: one or more machine learning methods in deep learning methods, support vector machines, adaboost, and random forest algorithms.

For example, when identifying the three-dimensional volume data of the fetal spine from the three-dimensional volume data of the second tissue to be measured by template matching, the prior data is used as a template to calculate the edge gradient, grayscale and other pixel features of the template, and traverse the The second three-dimensional volume data of the tissue to be measured is used to find the optimal solution with the smallest difference with the pixel feature of the template, so as to realize the identification of the fetal rib structure.

Another example: when the three-dimensional volume data of the fetal spine is identified from the three-dimensional volume data of the second tissue to be measured by the seed area growing method, the seed area is determined in the fetal spine area, and the pixels in the seed area are used as the seed pixels, and then According to the second pixel feature of the seed pixel, from the 3D volume data of the second tissue to be measured, the pixels conforming to the second pixel feature are merged into the seed region, and the newly added pixel is used as a new seed pixel to continue merging until no more pixels are found. Until the new pixels that meet the conditions, the three-dimensional volume data of the fetal spine is finally identified.

Here, certain prior data is pre-acquired in the three-dimensional volume data of the fetal spine as known information through a certain workflow of user interaction, and the difficulty of identifying the fetal spine is reduced through the pixel features of the known information.

Spine identification method 3. Determine at least two fourth candidate areas from the three-dimensional volume data of the second tissue to be measured, and obtain the volume data characteristics of the three-dimensional volume data of each fourth candidate area; The volume data characteristics of the candidate area, determine the fourth matching degree of each fourth candidate area and the fetal spine; determine the fourth candidate area with the highest fourth matching degree as the target area corresponding to the fetal spine; The three-dimensional volume data of the target area is used as the three-dimensional volume data of the fetal spine.

Here, the three-dimensional volume data of the second tissue to be measured may be received through the spine detection model, the spine detection model determines at least two fourth candidate regions from the three-dimensional volume data of the second tissue to be measured, and obtains each fourth candidate region The volume data characteristics of the three-dimensional volume data; the spine detection model determines the fourth matching degree between each fourth candidate region and the fetal spine according to the volume data characteristics of each fourth candidate region; the spine detection model determines the fourth matching degree The fourth highest candidate area is the target area corresponding to the fetal spine; the spine detection model uses the three-dimensional volume data of the target area corresponding to the fetal spine as the three-dimensional volume data of the fetal spine.

The algorithm adopted by the spine detection model can be a machine learning method. The spine detection model uses the three-dimensional volume data of the fetal spine as a training sample, and uses the machine learning method to learn the training samples, so as to train the spine detection model through the training samples. The trained spine detection model learns the volume data features of the fetal spine, wherein the volume data features may include: PCA features, LDA features, Harr features, texture features and other features.

When the spine detection model receives the 3D volume data of the second tissue to be measured, it recognizes the included 3D volume data of the fetal spine from the received 3D volume data according to the learned features of the volume data of the fetal spine.

In the training samples, the fetal spine is used as the target, and the target is calibrated. Wherein, the calibration can be performed by means of an ROI box containing the target, or by a mask (Mask) which accurately segments the target.

The algorithm adopted by the spine detection model can be an image segmentation algorithm, which performs binary segmentation on the 3D volume data input into the spine detection model, and performs operations such as morphology, contour extraction, and Unicom domain to obtain multiple fourth candidate regions. The volume data characteristics of each fourth candidate area determine the probability that each fourth candidate area is the fetal spine, select the fourth candidate area with the highest probability as the target area corresponding to the fetal spine, and the three-dimensional volume data of the selected target area is the fetus 3D volumetric data of the spine.

In practical applications, the image segmentation algorithm adopted by the spine detection model can also be: Level Set, Graph Cut, Snake, Random walker, Active Contour Model Algorithm, Active Shape Model Algorithm, One or more of the active appearance model algorithms, and image segmentation algorithms in deep learning such as FCN and UNet.

The algorithm adopted by the spine detection model may also be a template matching algorithm to establish a template of the three-dimensional volume data of the fetal spine. The spine detection model performs binary segmentation on the input 3D volume data, and performs operations such as morphology, contour extraction, and Unicom domain to obtain multiple fourth candidate areas, and traverses all the fourth candidate areas in the volume data according to the established template , and determine the similarity between all the fourth candidate regions and the template, that is, the fourth matching degree, select the fourth candidate region with the highest similarity as the target region corresponding to the fetal spine, and the three-dimensional volume data of the target area is the three-dimensional volume data of the fetal spine .

It should be noted that, in the embodiment of the present application, the specific identification manner of identifying the three-dimensional volume data of the fetal spine from the three-dimensional volume data of the second tissue to be measured is not limited in any way.

S604. Segment the 3D volume data of the fetal spine from the 3D volume data of the second tissue to be measured;

After identifying the three-dimensional volume data of the fetal spine, according to the position of the three-dimensional volume data of the fetal spine in the three-dimensional volume data of the second tissue to be measured, the three-dimensional volume data of the fetal spine are divided from the three-dimensional volume data of the second tissue to be measured come out.

In practical applications, S603 and S604 can be implemented simultaneously, and at the same time, the three-dimensional volume data of the fetal spine is recognized, and the three-dimensional volume data of the fetal spine are separated from the three-dimensional volume data of the second tissue to be measured.

S605. Identify the three-dimensional volume data of the lumbar spine and the three-dimensional volume data of the lumbar spine from the three-dimensional volume data of the fetal spine;

In S604, after separating the 3D volume data of the fetal spine from the 3D volume data of the second tissue to be measured, when identifying the 3D volume data of the lumbar spine from the 3D volume data of the fetal spine, at least the following two types of lumbar spine identification may be included: One of the ways:

Lumbar recognition method 1. Based on the shape characteristics of the vertebrae, locate the vertebrae of the fetal spine from the three-dimensional volume data of the fetal spine; display the three-dimensional image corresponding to the three-dimensional volume data of the fetal spine; The 3D image corresponding to the 3D volume data receives a sixth input operation; according to the sixth input operation, determine a reference vertebra and a vertebra identification corresponding to the reference vertebra; the reference vertebra is at least one vertebra in the fetal spine; based on The reference vertebra and the vertebra identification corresponding to the reference vertebra identify the three-dimensional volume data of the lumbar vertebra.

After the three-dimensional volume data of the fetal spine is determined, a three-dimensional image corresponding to the three-dimensional volume data of the fetal spine is displayed on the display. Based on tools such as a trackball and a touch screen, the user performs a sixth input operation on the 3D image corresponding to the 3D volume data of the spine displayed on the display to identify some vertebrae in the fetal spine and the identified vertebrae of the fetal spine logo.

After receiving the sixth input operation, the ultrasonic imaging device determines the vertebra identified by the user through the sixth input operation, takes the vertebra identified by the user as the reference vertebra, and uses the vertebra identification input by the user as the vertebra identification of the reference vertebra, and determines the The lumbar vertebrae in the fetal spine and the corresponding vertebral marks of the lumbar vertebrae.

When the ultrasonic imaging equipment determines the lumbar vertebra in the fetal spine and the rib mark corresponding to the lumbar vertebra according to the reference vertebra, it can use the gray histogram projection method, the contour extraction algorithm, the edge extraction algorithm, the Unicom domain method, the mass detection algorithm, and the template matching algorithm One or more methods among the image feature extraction algorithm and the morphological operation algorithm determine the vertebrae included in the fetal spine, and determine the lumbar vertebrae from the determined vertebrae included in the fetal spine.

Lumbar spine identification method 2: Identify the 3D volume data of the lumbar spine from the 3D volume data of the fetal spine based on the lumbar spine detection model for the lumbar spine. The lumbar spine detection model is obtained by training the sample volume data of the lumbar spine.

The algorithm adopted by the lumbar spine detection model can be a machine learning method. The lumbar spine detection model uses the three-dimensional data of the lumbar spine as a training sample, and uses the machine learning method to learn the training samples, so as to train the lumbar spine detection model through the training samples. The trained lumbar spine detection model learns the volume data features of the lumbar spine, wherein the volume data features may include: PCA features, LDA features, Harr features, texture features and other features.

When the lumbar spine detection model receives the three-dimensional volume data of the fetal spine, it identifies the included three-dimensional volume data of the lumbar spine from the received three-dimensional volume data according to the learned features of the volume data of the lumbar spine.

In the training samples, the lumbar spine is used as the target, and the target is calibrated. Wherein, the calibration can be performed by means of an ROI box containing the target, or by a mask (Mask) which accurately segments the target.

The algorithm adopted by the lumbar spine detection model can be an image segmentation algorithm, which performs binary segmentation on the 3D volume data input into the lumbar spine detection model, and performs operations such as morphology, contour extraction, and Unicom domain to obtain multiple candidate regions. The volume data features of the candidate area determine the probability that each candidate area is the lumbar spine, select the candidate area with the highest probability as the area corresponding to the lumbar spine, and the 3D volume data of the selected area is the 3D volume data of the lumbar spine.

In practical applications, the image segmentation algorithm adopted by the lumbar spine detection model can also be: Level Set, Graph Cut, Snake, Random walker, Active Contour Model Algorithm, Active Shape Model Algorithm, One or more of the active appearance model algorithms, and image segmentation algorithms in deep learning such as Fully Convolutional Networks (FCN) and UNet.

The algorithm adopted by the lumbar vertebra detection model may also be a template matching algorithm, which establishes a template of the three-dimensional volume data of the lumbar vertebra. The lumbar spine detection model performs binary segmentation on the input 3D volume data, and performs operations such as morphology, contour extraction, and Unicom domain to obtain multiple candidate areas. According to the established template, it traverses all candidate areas in the volume data and determines all The similarity between the candidate area and the template, the candidate area with the highest similarity is selected as the target area corresponding to the fetal rib or spine, and the three-dimensional volume data of the target area is the three-dimensional volume data of the lumbar spine.

The algorithm adopted by the lumbar spine detection model can also be: one or more methods such as image edge extraction, histogram image gray projection statistics, image contour extraction, morphological processing, threshold segmentation, and mass detection can directly calculate the fetal rib structure. Number of fetal ribs in 3D volume data.

In practical applications, the lumbar detection model and the first lumbar detection model may be a pass-through detection model.

S605. Render the 3D volume data of the conus medullaris and the 3D volume data of the lumbar spine to obtain a 3D ultrasound image of the vertebral structure;

Identify the three-dimensional volume data of the conus medullaris in the second tissue to be measured in S602, and after identifying the three-dimensional volume data of the lumbar spine in the second tissue to be measured in S605, perform the three-dimensional volume data of the conus medullaris and the three-dimensional volume data of the lumbar spine 3D rendering to obtain a 3D ultrasound image of the vertebral structure.

Here, when three-dimensional rendering is performed on the three-dimensional volume data of the conus medullaris and the three-dimensional volume data of the lumbar spine, all the data except the three-dimensional volume data of the conus medullaris and the three-dimensional volume data of the lumbar spine in the three-dimensional volume data of the second tissue to be measured are Empty, the three-dimensional volume data of the spinal cord and the three-dimensional volume data of the lumbar spine are rendered in stereo light perspective to form a three-dimensional ultrasound image of the vertebral structure.

It should be noted that there may be multiple ways of rendering, and this embodiment of the present application does not impose any limitation on the way of rendering.

S606. Mark the conus medullaris in the three-dimensional ultrasound image of the vertebral structure;

According to the position of the three-dimensional data of the conus medullaris identified in S502 in the three-dimensional volume data of the second tissue to be measured, determine the position of the three-dimensional data of the conus medullaris in the three-dimensional ultrasonic image of the fetal rib structure in S203, and according to the determined position of the spinal cord The position of the cone is used to mark the conus medullaris, and a three-dimensional ultrasound image of the marked vertebral structure is obtained.

In practical application, the distance between the conus medullaris and the lumbar vertebra is detected, and the detected distance is displayed on the display screen. And compare the detected distance with the set distance range, if the detected distance does not match the set distance range, an alarm will be issued.

S607. Output the marked three-dimensional ultrasonic image of the vertebral structure.

At this time, the content displayed on the monitor of the ultrasound imaging device is: a three-dimensional ultrasound image of the vertebral structure marked with the conus medullaris, enabling the user to visually view the position of the conus medullaris of the fetus and determine the distance between the conus medullaris and the lumbar spine .

The ultrasonic imaging method provided in the embodiment of the present application automatically or semi-automatically or semi-automatically recognizes, locates, and segments the three-dimensional volume data of the fetal spine after obtaining the three-dimensional volume data of the fetus, and performs automatic imaging on the three-dimensional VR map of the vertebral structure. Mark the conus medullaris on the VR map. This greatly simplifies the workflow of conus medullaris examination, liberates doctors from cumbersome and complicated manual operations, reduces the dependence on doctors' skills, and improves examination efficiency; and the stability and imaging quality of conus medullaris positioning results are compared with manual Are in a better state, reducing the rate of misdiagnosis and missed diagnosis.

In the related art, an all-manual method is still adopted in the ultrasonic fetal rib inspection and the position inspection of the conus medullaris. For example: After obtaining the ultrasound volume data of the fetal ribs, doctors often need to adjust the rotation and translation of the X, Y, and Z axes to adjust the orientation of the data, and manually adjust the size and position of the VOI (Volume of Interest). In order to better observe the fetal rib structure or the overall structure of the spine. For the fetal rib diagram shown in Figure 7, the area 701 corresponding to the rib structure needs to manually count the number or abnormalities of the fetal ribs to manually check whether the ribs are real or abnormal. If the rib A in the area 701 is abnormal, it is necessary to manually determine the rib A. Location.

If you want to obtain the transverse or coronal plane of a certain fetal rib, all fetal ribs or vertebrae, as shown in Figure 8, you need to use the Curved Multi Planar Reformation (CMPR) Anatomical trajectory 801 is drawn on the fetal rib or vertebra, parameters such as thickness are set, and the abnormality of the affected rib or vertebra can only be observed after obtaining the stretched transverse section or coronal plane 8(B) of the specified rib or vertebra. In the examination of the position of the conus medullaris, the doctor needs to manually and continuously adjust the sagittal plane of the spine to the optimal position, and then subjectively think that the conus medullaris has the best imaging effect. The position is relative to the lumbar spine. This method is more subjective and is easily affected by the imaging quality and viewing angle, resulting in inaccurate results.

It can be seen that in the related art, each manual operation in the inspection steps of fetal ribs or conus medullaris is complicated, time-consuming, and very dependent on the doctor's skills and experience. For example, in the steps of operating CMPR, manual tracing requires a high More importantly, the quality of the doctor's tracing directly affects the quality of the transverse or coronal imaging of the ribs or vertebrae, which makes the imaging effect and quality lack consistency and stability, thus affecting the diagnostic results .

The ultrasonic imaging method provided in this application can effectively help doctors to carry out auxiliary diagnosis of diseases, significantly improve work efficiency, and improve the quality of key diagnostic data obtained. As an example, the ultrasound imaging equipment applied to the ultrasound imaging method of the embodiment of the present application may be shown in FIG. Wherein, the modules shown in FIG. 10 may be located in the processor 105 in the ultrasound imaging device shown in FIG. 1 .

The volume data acquisition module 1001 is used for acquiring three-dimensional volume data of ultrasound fetal ribs, fetal spine (also called spine) and spinal cord.

When the transmitting circuit 101 in Fig. 1 sends a group of delayed focused pulses to the probe 102, the probe 102 transmits ultrasonic waves to the tissue to be measured, and after a certain delay, receives the ultrasonic waves with tissue information reflected from the tissue to be measured echo, and convert this ultrasonic echo back into an electrical signal. The receiving circuit 103 receives these electrical signals, and sends these ultrasonic echo signals to the beamforming module 104 . The ultrasonic echo signal is completed in the beamforming module 104 with focus delay, weighting and channel summation, and then processed by the processor 105 to obtain three-dimensional volume data including fetal rib structure and spine.

The recognition module 1002 is used for the recognition, location and segmentation of the ultrasound rib structure or spine, and the recognition, location and segmentation of the conus medullaris.

Among them, the identification and positioning methods of fetal rib structure (also called rib structure) can be divided into manual, semi-automatic and automatic. and centerline. Fetal ribs are identified as follows:

1. Manually obtain the position of the fetal rib structure

The user uses trackball, touch screen and other tools to inform the system of the position of the rib structure in space through a certain workflow on the rib structure in the volume data, such as selecting marker points and drawing marker lines. For example, selecting the position of each fetal rib The endpoints or intermittently take some points on the border of the rib structure, roughly draw the midline of the rib structure or trace the boundary line of the rib structure, etc.

2. Method for automatic identification of fetal rib structure

Using machine learning methods to learn the characteristics or rules that can distinguish fetal ribs from other non-rib structural tissues in the database, and then locate and identify fetal rib structures in other volume data according to the learned features or rules, first use multiple fetal ribs The volume data and the corresponding calibration results build a database. Among them, the calibration result can be set according to the actual task needs, and it can be an ROI frame containing the target, or a Mask (mask) for accurately segmenting the target.

It should be noted that if each rib or spine of the fetus is regarded as a different category to identify and locate the target instead of identifying the structure composed of all ribs and spine as an overall target, you also need to specify each ROI box or Mask The category of ribs or spines, that is, it is necessary to mark all fetal ribs and spines into different categories, which is transformed into a multi-target recognition and positioning problem. After building the database, design a machine learning algorithm to learn the characteristics or rules of the fetal rib or spine area and non-fetal rib or non-spine area in the fetal rib and spine database to realize the positioning and identification of fetal ribs and spine in volumetric data .

Methods for identifying fetal rib structure include but are not limited to the following:

Method 1. The method based on the sliding window: first, feature extraction is performed on the area in the sliding window. The extracted features can be PCA features, LDA features, Harr features, texture features, etc., or a deep neural network can be used for feature extraction, and then Match the extracted features with the database, classify with KNN, SVM, random forest, neural network and other discriminators, determine whether the current sliding window is an area of interest and obtain its corresponding category.

Method 2. Bounding-Box method recognition based on deep learning. The common form is: by stacking the basic convolutional layer and the fully connected layer to learn the characteristics of the constructed database and return the parameters. For the input 3D volume data , the Bounding-Box of the corresponding region of interest can be directly regressed through the network, and the category of the organizational structure in the region of interest can be obtained at the same time. The algorithms used in the network model include R-CNN, Fast R-CNN, Faster-RCNN, SSD , YOLO, etc.

Method 3. An end-to-end semantic segmentation network method based on deep learning. This type of method is similar to the Bounding-Box structure of method 2 based on deep learning. The difference is that the fully connected layer is removed and an upsampling or deconvolution layer is added. Make the size of the input and output the same, so as to directly obtain the region of interest of the input image and its corresponding category. Common networks include FCN, U-Net, Mask R-CNN, etc.

Method 4: Use method 1, method 2 or method 3 to locate the target, and then use a classifier to classify and judge the target of the positioning result. The class judgment method can be as follows: first, feature extraction is performed on the target ROI or Mask, and the extracted feature extraction method can be PCA feature, LDA feature, Harr feature, texture feature, etc., or a deep neural network can be used for feature extraction, and then the extracted feature The obtained features are matched with the database, and the discriminators such as KNN, SVM, random forest, and neural network are used for classification.

The method for identifying the fetal rib structure can be an image segmentation algorithm to accurately segment the fetal rib structure in the body data. Here, the fetal rib structure usually appears as a hyperechoic arc-shaped object with high gray value, and the rib structure can be segmented out by an image segmentation algorithm.

For example, first perform binary segmentation on the volume data, perform some necessary operations such as morphology, contour extraction, and connected domains to obtain many candidate areas, and then judge the area of each candidate area according to its shape, aspect ratio and other characteristics. The probability of the rib structure, the region with the highest probability is selected as the rib structure region. Other image segmentation algorithms can also be used, such as one of Level Set, Graph Cut, Snake, Random walker, Active Contour Model Algorithm, Active Shape Model Algorithm, and Active Appearance Model Algorithm Or more and some image segmentation algorithms in deep learning, such as FCN, UNet and so on.

As another example, the template matching method can also be used to detect the fetal rib structure in the volume data. For example, the shape of the fetal rib structure is relatively special. Some data of the fetal rib structure can be collected in advance to create a template, and all possible patterns in the volume data can be traversed during detection. The area of the template is similarly matched with the template, and the area with the highest similarity is selected as the target area.

3. The method of semi-automatically obtaining the fetal rib position:

Users are required to use tools such as trackballs and touch screens to pre-acquire certain prior data in the fetal rib structure volume data as known information or pre-set certain parameters or conditions through a certain user interaction workflow to help reduce the recognition effort. Difficulty, etc., and then combine one or more image processing methods in template matching algorithm, image feature extraction algorithm, edge extraction algorithm, morphological operation algorithm, graphcut algorithm, grabcut algorithm, level set method, active contour model algorithm, active shape One or more image segmentation algorithms in model algorithm, seed region growth method, region segmentation and merging method, one or more machine learning methods in deep learning method, support vector machine, adaboost, random forest algorithm to realize recognition and Determine the location of fetal rib structures in volumetric data.

For example, when using the method of template matching, it is necessary to obtain a sample body data in advance, and a trackball, touch screen, etc. can be used to extract data containing part or all of the fetal rib structure information to make a template, such as using a rectangular frame to intercept one or more The section contains the image of the fetal rib structure as a template, and then calculates the edge gradient, grayscale and other information of the template, and then uses the template to traverse the entire volume data to find the optimal solution with the smallest gap with the template information to realize the recognition of the fetal rib.

When the graphcut algorithm is used to segment the fetal rib region, it is also necessary to specify the pixel points representing the target region and the background region respectively in the fetal rib structure and non-fetal rib structure regions as seed points. Point and draw lines inside and outside the structure to specify different seed points; the grabcut algorithm needs to manually draw a frame to frame the rib structure target area with segmentation or draw some respectively in the rib structure area and non-rib structure area to specify the target area Lines of different colors from pixels in the background area can get a more perfect segmentation result.

Segmentation methods such as level set and active contour model require user interaction to give the initial contour curve, and then perform curve evolution according to functional energy minimization to approach the boundary of the fetal rib to realize the segmentation of the fetal rib; the seed region growth method also requires the user to provide A group of seed pixels representing different growth areas, and then merge the matching pixels in the field of seed pixels into the growth area represented by the seed pixels, and continue to merge the newly added pixels as new seed pixels until no matching pixels are found. Until the new pixel, the fetal rib structure is finally segmented. Alternatively, the user can reduce the recognition difficulty by narrowing the detection range by setting a rectangular frame through an interactive workflow or specifying some landmarks of the rib structure in advance.

Through the above method, the position of the fetal rib structure in the volume data can be identified.

In practical applications, the above method of identifying the fetal rib structure is also applicable to the identification of the spine.

It should be noted that when constructing a machine learning database and a template matching template, the sample can be the overall structure including the spine and all 12 ribs as a labeled sample or template, or the 12 ribs and spine can be regarded as different categories Multi-target labeling and recognition or template matching.

The identification and localization of the conus medullaris needs to automatically identify, segment the conus medullaris and determine its exact position in the spinal cord body data. The method is as follows:

Machine learning method is used to learn the characteristics or rules that can distinguish the conus medullaris from other non-conus medullaris in the database, and then locate and identify the conus medullaris in other volume data according to the learned features or rules. First, multiple spinal cord data are used The sampled multi-frame pictures and the corresponding calibration results build a database. Among them, the calibration result can be set according to the actual task needs, it can be the ROI box containing the target, or a Mask (mask) that can accurately segment the target; after building the database, design a machine learning algorithm to learn the conus medullaris database In this method, the characteristics or rules of the conus medullaris and non-conus medullaris can be distinguished to realize the location and identification of the conus medullaris in the volume data.

Commonly used methods include but are not limited to the following methods:

Method 1. The method based on the sliding window: first, feature extraction is performed on the area in the sliding window. The extracted feature method can be PCA feature, LDA feature, Harr feature, texture feature, etc., or a deep neural network can be used for feature extraction. Then match the extracted features with the database, classify with KNN, SVM, random forest, neural network and other discriminators, determine whether the current sliding window is an area of interest and obtain its corresponding category.

Method 2. Bounding-Box method detection and recognition based on deep learning. The common form is: by stacking the basic convolutional layer and the fully connected layer to learn the characteristics of the constructed database and return the parameters. For the input 3D volume data, you can The Bounding-Box of the corresponding region of interest is directly regressed through the network, and the category of the tissue structure in the region of interest is obtained at the same time. The methods that can be used for the network model include R-CNN, Fast R-CNN, Faster-RCNN, SSD, YOLO, etc.

Method 3. An end-to-end semantic segmentation network method based on deep learning. This type of method is similar to the second Bounding-Box structure based on deep learning. The difference is that the fully connected layer is removed and an upsampling or deconvolution layer is added. To make the input and output the same size, so as to directly obtain the region of interest of the input image and its corresponding category, common networks include FCN, U-Net, Mask R-CNN, etc.

Method 4: Only use method 1, method 2 or method 3 to locate the target, and then use a classifier to classify and judge the target of the positioning result. The classification judgment method can be as follows: first, feature extraction is performed on the target ROI or Mask, and the extracted features can be PCA features, LDA features, Harr features, texture features, etc., or a deep neural network can be used for feature extraction, and then the extracted features The features are matched with the database, and the discriminators such as KNN, SVM, random forest, and neural network are used for classification.

The method for identifying the conus medullaris can also be an image segmentation algorithm to accurately segment the conus medullaris in volume data. Here, the conus medullaris usually appears as a hyperechoic arc-shaped target with a high gray value, and the conus medullaris can be segmented by image segmentation algorithms.

For example, first perform binary segmentation on the volume data, perform some necessary operations such as morphology, contour extraction, and connected domains to obtain many candidate areas, and then judge the area of each candidate area according to its shape, aspect ratio and other characteristics. The probability of the conus medullaris, select a region with the highest probability as the conus medullaris region. Other image segmentation algorithms can also be used, such as one of Level Set, Graph Cut, Snake, Random walker, Active Contour Model Algorithm, Active Shape Model Algorithm, and Active Appearance Model Algorithm Or more and some image segmentation algorithms in deep learning, such as FCN, UNet and so on.

As another example, template matching method can also be used to detect the conus medullaris in the volume data. For example, the shape of the conus medullaris of the fetus is quite special. Some data of the conus medullaris can be collected in advance to create a template, and all possible conus medullaris in the volume data can be traversed during detection. The area is similarly matched with the template, and the area with the highest similarity is selected as the target area.

The accurate position of the fetal conus medullaris in the volume data can be automatically located through the above method.

The statistics module 1003 is used for semi-automatic or automatic statistics of the number of ultrasonic fetal ribs and vertebrae.

After the identification module 1002 determines the fetal rib structure or the position of the spine in the volume data, the statistical module 1003 identifies and marks the 12 fetal ribs, spine, and lumbar vertebrae by semi-automatic or automatic methods, calculates the number of fetal ribs and compares the number of ribs , spine, and lumbar vertebrae are displayed on the display module 1005.

It should be noted that the identification, position, and name marking of the lumbar vertebrae are used as a reference for the position of the end of the conus medullaris. Clinically, the specific vertebral body level corresponding to the end of the conus medullaris is often used to judge the position of the end of the conus medullaris and its location. Ascending rules, for example, the end of the conus medullaris in normal adults is located at the L1-2 level of the lumbar spine. The calculation and labeling of the number and position of the lumbar vertebrae are not used as a judgment index for congenital disease examination, but only as a reference and measurement index to indicate the specific position of the end of the fetal conus medullaris.

1. Quantity statistics of semi-automatic fetal rib bodies:

Combined with the information provided by the user through a certain manual operation workflow, on the basis of the fetal rib structure data segmented by the identification module 1002, methods including but not limited to the following are used: gray histogram projection method, contour extraction algorithm, edge extraction algorithm, Unicom One or more methods in field method, mass detection algorithm, template matching algorithm, image feature extraction algorithm, and morphological operation algorithm are used to count the number of fetal ribs or vertebral bodies, and mark the name of each rib or vertebral body.

For example, first, select one or more ribs or vertebral structures in the segmented fetal rib or vertebral body data through tools such as trackball and touch screen, such as pointing a point on T3 and T9 ribs or drawing T3 and T9 respectively. The centerline of T9 ribs, that is, the two ribs are marked as known, and then the result of rib volume data segmentation and the fixed arrangement sequence relationship and spatial position relationship between T1 to T12 ribs, such as in the case where the rib structure is not missing The ribs above the lower T2 ribs must be T1 and the lower ribs must be T3, and the number statistics and name marking are performed in combination with the above-mentioned algorithm.

For example, a certain coronal plane of volume data segmented by the histogram grayscale projection method is projected along the Y axis, and the number and position of fetal ribs are obtained by calculating the number of pixel peaks and the position at the peak of the pixel statistical map, respectively. Calculate the distance between the general area information at the points and lines drawn and the obtained position information. The ribs with the closest distances are marked as T3 and T9 ribs, and all other ribs are marked in turn or only the ribs between T3 and T9 are marked.

Similarly, the position of each segmented rib can also be determined according to the method of finding the contour, and the number of ribs can be calculated by calculating the outermost surrounding contour, that is, the number of the largest contour. Of course, some image preprocessing work may be required to remove some irrelevant structures. Reducing interference, the above functions can also be achieved by calculating the number and position of specific contours. Cluster detection, template matching, etc. are similar and will not be repeated here.

2. The method of automatically counting fetal ribs or vertebral bodies

After the fetal rib structure or spine volume data is segmented, the known spatial positional relationship between the fetal ribs can be used, such as the arrangement sequence of the ribs (T1-T12) shown in Figure 11, or some other known The prior knowledge combined with image processing methods include but not limited to one or more methods such as image edge extraction, histogram image gray projection statistics, image contour extraction, morphological processing, threshold segmentation, clump detection, etc. Number of fetal ribs in ex vivo data. Here, in the arrangement order of the ribs shown in FIG. 11 , only T3, T6, T9, and T12 are marked, that is, only the third rib, the sixth rib, the ninth rib, and the twelfth rib are marked.

For example, the fetal rib structure or the coronal section image of the spine extracted by machine learning or segmentation algorithm segmentation in the identification module 1002 is first subjected to threshold segmentation, and then the histogram of the gray value is performed along the direction perpendicular to the spine or parallel to the rib Statistical projection, the ordinate is the image height, the abscissa is the number of pixels, and then set the threshold to calculate the peak number of the statistical graph to get the number and position of the fetal ribs and then mark them. Similarly, the position of the spine can also be marked.

As another example, the number of ribs can be directly obtained by combining contour extraction and mass detection.

Machine learning methods can also be used to take 12 different ribs and spines as a recognition target to build a learning database. The learning library can distinguish the characteristics or laws of each specific rib from other ribs or spines to learn the characteristics and patterns of each different rib and spine. The characteristics or rules of the spine Locate and identify the ribs and spine in the fetal rib structure in other volume data. The calibration results can be set according to the actual task needs, which can be the ROI box containing the target, or the target. Accurately segmented Mask (mask); here, each rib or spine of the fetus is regarded as a target of different categories to identify and locate. It is necessary to specify the category of each ROI box or the rib or vertebra of the Mask, which is a multi-target recognition and positioning problem. . After building the database, design a machine learning algorithm to learn the characteristics or rules that can distinguish the fetal rib area and non-fetal rib area in the fetal rib database to realize the positioning and identification of fetal ribs in the volume data.

Commonly used methods are similar to the recognition module 1002, such as: Method 1, the traditional method of feature extraction based on sliding windows and classification by discriminators. Method 2: Bounding-Box method detection and recognition method based on deep learning. Method 3. An end-to-end semantic segmentation network method based on deep learning. Method 4: Only use method 1, method 2 or method to locate the target, and then use a classifier to classify and judge the target of the positioning result.

The template matching method can also be used to detect different ribs and spines in the fetal rib structure data. For example, according to each different rib and spine of the fetus, different templates are established, and the fetal rib structures extracted by traversing the segmentation are traversed during detection. All possible areas are similarly matched with the template, and the area with the highest similarity is selected as the target area, and then marked and counted according to the matching results.

By including but not limited to a combination of one or more of the above methods, the rib structures in the identified fetal rib structures can be counted and the number of ribs can be marked and counted.

It should be noted that the method for counting the number of vertebral bodies in the spine by the statistical module 1003 is the same as the method for counting the number of fetal ribs in the fetal rib structure, so that the number of vertebral bodies included in the spine and the number of each vertebral body can be counted. mark, so as to determine the lumbar spine.

The imaging module 100 is used for automatic imaging of fetal rib structure.

During the automatic imaging of the fetal rib structure, the straightening algorithm is used to straighten and reconstruct the rib structure data, and then the required volume data section is identified and positioned for imaging. As shown in Figure 11, the main content of the automatic imaging includes automatic imaging of the coronal plane of ribs or ribs being straightened, automatic imaging of the cross-section of each rib or ribs being straightened, automatic imaging of the cross-section of manually selected ribs, Automatic imaging of three-dimensional rib skeleton extraction and automatic imaging of vertebral VR images marked with the position of the end of the conus medullaris, etc.

The volume data of the ribs and the volume data of the spine need to be straightened and reconstructed to obtain complete coronal and transverse planes. The straightening of the straightening object is divided into two parts: the extraction of the longitudinal axis and the straightening and reconstruction, and the straightening object includes ribs and spine.

The method for extracting the longitudinal axis of the straightened object may be: a tracking-based longitudinal axis extraction algorithm, a model-based multi-scale longitudinal axis extraction algorithm, a morphology-based longitudinal axis extraction method, a region growing-based centerline extraction method, a three-dimensional based The method of geometric moments and the method of locating the centerline using machine learning, etc.

The tracking-based longitudinal axis extraction algorithm is a semi-automatic algorithm. Based on the initial key points and end points provided by user interaction, a section perpendicular to the tracking direction is generated during the tracking process, and the section is accurately calculated using the maximum likelihood method and the centroid method. The center point of the middle rib or spine. After the tracking is completed, interpolation and fitting of the center point sequence is performed to obtain the longitudinal axis. At this time, equally spaced sampling is performed on the longitudinal axis to generate a sequence of equally spaced rib or spine sections perpendicular to the direction of the ribs or spine, and finally the equally spaced sections are reconstructed to achieve the purpose of straightening the ribs or spine.

The model-based multi-scale longitudinal axis extraction algorithm approximates the local rib or spine as a tubular structure, and uses the center of gravity of the tubular structure obtained by calculating the geometric moments as the center of the local rib or spine, through the Hessian corresponding to a voxel under multi-scale Gaussian filtering The eigenvalue of the Hessian matrix is analyzed, the local rib or vertebral structure is enhanced, and the direction of the longitudinal axis is estimated according to the eigenvector corresponding to the minimum eigenvalue of the Hessian matrix. After the longitudinal axis is obtained, the ribs and vertebrae can be straightened and reconstructed. First, the longitudinal axis of the ribs is sampled at equal intervals to obtain the center points at equal intervals. Stack them together to obtain rib section views or 3D reconstruction views from different angles to achieve the purpose of straightening.

The imaging of the coronal plane of the rib, the imaging of the cross section of the rib, the imaging of the rib skeleton and the labeling of the conus medullaris will be described respectively below. Wherein, the object to be imaged in the imaging of the coronal plane of the rib and the imaging of the cross section of the rib may be a rib or a straightened rib.

First, the imaging of the coronal plane of the rib

The display of the coronal plane of the rib can only be obtained after straightening the spine and the rib at the same time. As shown in FIG. After the straightening method straightens the rib body data in the two directions indicated by the arrows, the longitudinal axes of all the extracted ribs and spine are almost all located on the same plane, directly by solving mathematical equations or similar least squares method or Hough transform The method of isoplane fitting determines the plane, and according to the volume data imaging of the plane, the required coronal plane of the fetal rib structure is obtained.

Second, the imaging of the cross section of the rib

The display of the cross-section of the ribs needs to straighten the ribs along the long axis 1202 in FIG. 12 . In order to accurately find and determine the cross-sections of all the ribs, here the spine is still straightened along the direction of the arrow 1202 in FIG. 12 . Use two straight lines to determine a plane to determine the cross-section of each rib. Among them, the transverse plane must be perpendicular to the longitudinal axis of the spine and coplanar with the longitudinal axis of the ribs, so that the transverse plane without ribs can be uniquely determined.

The display of rib cross-sections is divided into two display modes: the display of all rib cross-sections shown in Figure 13 and the automatic display of manually selected specified rib cross-sections shown in Figure 14.

Third, the imaging of the rib skeleton

The structure of each rib T1-T12 and the centerline of the spine are identified through the identification module 1002 or the statistics module 1003, and the identified ribs are marked.

Firstly, new volume data is reconstructed from the volume data of the rib and spine structures obtained by recognition and segmentation. For example, the gray level of the fetal rib structure is 1, and the gray value of the non-fetal rib structure is 0. Then, use the method of volume rendering or surface rendering to render the reconstructed volume data, and obtain the three-dimensional skeleton rendering diagram of fetal rib and vertebra structure as shown in FIG. 15 .

In practical applications, the cross-section and three-dimensional skeleton of each straightened rib can be displayed on the same display interface.

Fourth, the labeling of the conus medullaris

The specific position of the end of the conus medullaris on the sagittal plane (two-dimensional coordinates) located by the identification module 02, and combined with the specific position of the sagittal plane in the volume data to calculate the three-dimensional geometric coordinate point of the end of the conus medullaris in the volume data Finally, according to the three-dimensional coordinate points and the posture of the fetal spine in the body data identified by the positioning module (such as the obtained centerline of the spine), a plane or line that passes through the end point of the conus medullaris and is perpendicular to the spine is calculated. By mapping this plane or straight line to the stereoscopic VR map of the vertebral body structure, the relative position of the end of the conus medullaris relative to the fetal lumbar vertebrae can be intuitively expressed, and this result (such as between L1 and 2) is calculated and marked in VR on the graph.

In practical applications, fetal ribs, lumbar vertebrae and conus medullaris can be detected simultaneously, and ribs and conus medullaris are marked on VR images including fetal rib structure and vertebral body structure. As shown in FIG. 16 , a straight line 1701 indicates the position of the vertebra corresponding to the specific position of the conus medullaris located.

The present application proposes an ultrasonic detection method for abnormal number of fetal ribs and abnormal position of conus medullaris. After obtaining the fetal three-dimensional volume data, automatically or semi-automatically (manually designate 1 or 2 ribs), identify, locate and segment the fetal rib structure and automatically locate the exact position of the fetal conus medullaris, and perform automatic/semi-automatic statistics of the number of fetal ribs , as well as (3D) extraction of the rib skeleton from the volume data and automatic imaging of the cross-section, coronal plane and stereo VR map of all ribs or specified ribs after being straightened, marking the conus medullaris in the lumbar spine on the stereo VR map specific location.

This method greatly simplifies the workflow of fetal rib inspection and conus medullaris position inspection, frees doctors from cumbersome and complicated manual operations, and puts more energy into the diagnosis of diseases, reducing the need for doctors' skills. Dependence, improve inspection efficiency; and the stability and imaging quality of the rib number statistics and conus medullaris location results are in a better state than manual, reducing the rate of misdiagnosis and missed diagnosis.

In one embodiment, an ultrasound imaging method is also provided. The ultrasonic imaging method can be applied to the aforementioned ultrasonic imaging device. In this method, the three-dimensional volume data of the fetal ribs can be automatically or semi-automatically identified from the three-dimensional volume data of the fetus based on the characteristics of the fetal ribs, and then automatically or semi-automatically obtained based on the identified three-dimensional volume data of the fetal ribs. Sectional images of clinical value, such as coronal images, cross-sectional images, longitudinal cross-sectional images of ribs, etc.

For example, in this embodiment, the three-dimensional volume data of the fetus can be obtained first. The three-dimensional volume data of the fetus can be obtained by real-time scanning of the ultrasound imaging equipment. For example, the ultrasound imaging equipment transmits ultrasound to the fetus through an ultrasound probe and receives ultrasound echoes to obtain ultrasound echo signals. The processor of the ultrasound equipment processes the ultrasound echo signals. processing to obtain the three-dimensional volume data of the fetus. The three-dimensional volume data of the fetus can also be pre-acquired and stored by the ultrasound imaging equipment, and read in for processing when it is necessary to obtain a section image or three-dimensional image of the fetal rib.

Then, the processor may identify the three-dimensional volume data of the fetal ribs from the three-dimensional volume data of the fetus based on the features of the fetal ribs. Here, the fetal ribs will show specific features on the ultrasound image due to their own characteristics. Therefore, the processor can identify the three-dimensional volume data of the fetal ribs based on the image features of the fetal ribs (ie, the features exhibited by the image data), for example, the grayscale or the mean value, variance, distribution characteristics, texture, Morphological features, etc.

Then, according to the identified three-dimensional volume data of fetal ribs, the processor can obtain at least two ribs in the three-dimensional volume data of fetal ribs that are parallel to or coincident with the arrangement planes of multiple fetal ribs in the three-dimensional volume data of fetal ribs. The first plane (for example, coronal plane) or the first curved surface and/or the first rib that passes through at least one rib in the three-dimensional volume data of fetal ribs and intersects with the arrangement plane of multiple fetal ribs in the three-dimensional volume data of fetal ribs second plane (for example, cross-section) or second curved surface, and then obtain an image on the first plane or the first curved surface and/or obtain an image on the second plane or the second curved surface according to the identified three-dimensional volume data of the fetal rib.

After obtaining the image on the first plane or the first curved surface and/or the image on the second plane or the second curved surface, the processor can display the image on the first plane or the first curved surface as a two-dimensional image and/or Or display the image on a second plane or a second curved surface as a 2D image.

In this embodiment, the method may further include: obtaining a three-dimensional ultrasonic image of the fetal rib according to the identified three-dimensional volume data of the fetal rib, and displaying the three-dimensional ultrasonic image of the fetal rib.

In this embodiment, based on the first rib detection model, the three-dimensional volume data of the fetal spine and the three-dimensional volume data of the fetal ribs can be automatically identified from the three-dimensional volume data of the first tissue to be measured.

In this embodiment, the three-dimensional volume data of the fetal ribs may also be identified from the three-dimensional volume data of the fetus based on the features of the fetal ribs by using a template matching method. For example, at least two first candidate areas may be determined from the three-dimensional volume data of the fetus, the volume data features of the three-dimensional volume data of each first candidate area may be obtained, and each candidate area may be determined according to the volume data features of each first candidate area. a first matching degree between the first candidate region and the fetal rib, and determining that the first candidate region with the highest first matching degree is the target region corresponding to the fetal rib, and using the three-dimensional volume data of the target region corresponding to the fetal rib as the three-dimensional data of the fetal rib volume data.

In this embodiment, when identifying the three-dimensional volume data of fetal ribs, it can also be performed based on the input operation of the user or other equipment connected to the ultrasound imaging equipment through a wired or wireless network, that is, it can be performed semi-automatically. For example, the 3D ultrasound image corresponding to the 3D volume data of the fetus can be displayed, and based on the 3D ultrasound image corresponding to the 3D volume data of the fetus, the first input operation is received, the marker point corresponding to the first input operation is determined, and the coordinates of the marker point are obtained from The three-dimensional volume data of the ribs of the fetus are identified in the three-dimensional volume data of the fetus.

In this embodiment, the fetal ribs and the overall structure of the fetal spine may also be identified first, and then the fetal ribs may be identified therefrom. For example, the 3D volume data of the fetal rib structure can be identified from the 3D volume data of the fetus, wherein the fetal rib structure includes fetal ribs and fetal spine, and then the 3D volume data of the fetal ribs can be identified from the 3D volume data of the fetal rib structure.

In this embodiment, the ribs may also be straightened, and then an image on the first plane or the first curved surface and/or an image on the second plane may be obtained based on the three-dimensional volume data of the straightened ribs. For example, the 3D volume data of the identified fetal ribs can be straightened to obtain the 3D volume data of the straightened ribs, and then according to the 3D volume data of the straightened ribs, at least two ribs in the 3D volume data of the straightened ribs can be obtained and combined with the straightened ribs. A first plane or a first curved surface parallel to or coincident with the arrangement planes of multiple fetal ribs in the rib three-dimensional volume data, and/or, at least one rib in the straightened rib three-dimensional volume data is obtained and combined with the straightened rib three-dimensional volume data The second plane intersects the arrangement planes of multiple fetal ribs, and according to the straightened rib three-dimensional volume data, an image on the first plane or the first curved surface and/or an image on the second plane is obtained.

In this embodiment, the fetal spine may also be identified, and after both the fetal ribs and the fetal spine are straightened, an image on the first plane and/or the second plane is obtained according to the straightened three-dimensional volume data. For example, based on the characteristics of the fetal spine, the three-dimensional volume data of the fetal spine can be identified from the three-dimensional volume data of the fetus. The identified fetal ribs and the identified fetal spine are usually connected to each other. Called the fetal rib structure. In this embodiment, the 3D volume data of the fetal rib structure can be straightened, that is, both the 3D volume data of the fetal ribs and the 3D volume data of the fetal spine can be straightened to obtain the 3D volume data of the straightened rib structure. Correspondingly, the straightening The rib structure 3D volume data will include straightened rib 3D volume data and straightened spine 3D volume data. Then, according to the three-dimensional volume data of the straightened rib structure, obtain the first plane passing through the straightened ribs and the straightened spine and/or obtain the second plane passing through at least one straightened rib and intersecting the straightened spine, and according to the Straighten the three-dimensional volume data of the rib structure to obtain an image on the first plane, and/or obtain an image on the second plane according to the straightened three-dimensional volume data of the rib structure.

In this embodiment, when the image on the first plane is obtained, the image on the first plane can be obtained based on data within a certain thickness range, so that the image on the first plane can display more information. For example, three-dimensional volume data within a predetermined thickness range in a direction perpendicular to the first plane may be obtained from the straightened rib three-dimensional volume data and/or straightened spine three-dimensional volume data, and obtained according to the three-dimensional volume data within the predetermined thickness range image on the first plane.

The three-dimensional volume data within the predetermined thickness range can be manually determined by the user. For example, the user can directly draw a predetermined thickness range on the displayed three-dimensional image or two-dimensional section image of the straightened three-dimensional volume data through the input device, or can also use the input device to draw a predetermined thickness range on the displayed straightened three-dimensional volume data. Determine the predetermined plane (for example, the plane where the upper and lower edges of the vertebral arch or vertebral body are located, or other planes according to actual needs, etc.) on the three-dimensional image or the two-dimensional section image, and then the area between the upper and lower edge planes Predetermined thickness ranges, etc.

In this embodiment, the three-dimensional volume data within the predetermined thickness range may include vertebral arches and/or vertebral bodies of the fetal spine. At this time, also, the predetermined thickness range can be manually set so that it includes the vertebral arch and/or vertebral body of the fetal spine. Alternatively, the predetermined thickness range can also be determined automatically or semi-automatically. For example, based on the characteristics of the vertebral arches and/or vertebral bodies of the spine, the vertebral arches and/or vertebral bodies can be identified from the straightened spine 3D volume data, and the 3D volume data within the predetermined thickness range can be determined so that the 3D volumetric data within the predetermined thickness range The volume data includes the identified vertebral arch and/or vertebral body, for example, three planes of the upper and lower edges can be fitted according to the identified vertebral arch and/or the upper and lower edges of the vertebral body, and any two planes in the three planes The range between them is the aforementioned predetermined thickness range.

In this embodiment, the method of identifying the vertebral arch and/or vertebral body from the three-dimensional volume data of the straightened spine based on the characteristics of the vertebral arch and/or vertebral body of the spine can refer to the identification of specific tissues from the three-dimensional volume data in the previous embodiments. The method of the structure can use a method similar to it, which will not be described in detail here.

In this embodiment, the image on the first plane can be obtained according to the three-dimensional volume data within a predetermined thickness range in various suitable ways. For example, the three-dimensional volume data within a predetermined thickness range may be weighted in a direction perpendicular to the first plane to obtain an image on the first plane, and so on. For example, the weighted average of all voxels on a path perpendicular to the first plane within the thickness range can be used to obtain the value of the pixel corresponding to the path on the first plane in this direction, and similarly obtain all voxels on the first plane. The value of the pixel point can obtain the image on the first plane.

In this embodiment, multiple suitable methods may be used to straighten the identified three-dimensional volume data of the ribs of the fetus to obtain straightened rib three-dimensional volume data. For example, the longitudinal axis of the 3D volume data of the identified fetal ribs can be determined, and the 3D volume data of the identified fetal ribs can be sampled according to the longitudinal axis to obtain a slice sequence, and then the slice sequence can be reconstructed along a straight line to obtain a drawn 3D volumetric data of straight ribs.

In this embodiment, similarly, various suitable methods may be used to straighten the identified three-dimensional volume data of the fetal spine to obtain straightened three-dimensional volume data of the spine. For example, the longitudinal axis of the identified three-dimensional volume data of the fetal spine can be determined, and the identified three-dimensional volume data of the fetal spine can be sampled according to the longitudinal axis to obtain a slice sequence, and the slice sequence can be reconstructed along a straight line to obtain a straight line. Spine 3D volume data.

In this embodiment, the number of fetal ribs can also be automatically determined according to the recognized three-dimensional volume data of fetal ribs, and the number of fetal ribs can be displayed. The number of fetal ribs can be displayed with numbers or other suitable symbols.

In this embodiment, the fetal ribs can also be marked according to the recognized three-dimensional volume data of the fetal ribs, and the marks of the fetal ribs can be displayed. The marks may be various suitable marks, such as the aforementioned T1, T2 and other marks, and may also be other suitable words, numbers, symbols, colors and the like.

In this embodiment, the specific solutions of the various steps involved can refer to the methods in the foregoing embodiments or be the same as or similar to the similar steps in the foregoing embodiments, and will not be described in detail here.

In one embodiment, an ultrasonic imaging method is provided, which can be applied to the aforementioned ultrasonic imaging device. The method may include: acquiring the three-dimensional volume data of the fetus, identifying the conus medullaris region from the three-dimensional volume data of the fetus based on the characteristics of the conus medullaris, determining the position of the conus medullaris region according to the identified conus medullaris region, and displaying the conus medullaris region. Location of the conus medullaris. Here, various suitable ways can be used to display the position of the conus medullaris region, for example, by suitable symbols, colored areas, text, arrows, geometric shapes and the like.

In this embodiment, the determined position of the conus medullaris region may be the position of the end of the conus medullaris. For example, the position of the end of the conus medullaris can be determined based on the identified region of the conus medullaris, and the position of the end of the conus medullaris can be displayed. The location of the end of the conus medullaris can be displayed in a number of suitable ways, eg, a suitable compound, color, dot, line, arrow, number, distance from a suitable reference location, and the like.

In this embodiment, the region of the conus medullaris can be identified through the method of target matching. For example, at least two second candidate regions may be determined from the three-dimensional volume data of the fetus, the volume data characteristics of the three-dimensional volume data of each second candidate region are obtained, and each second candidate region is determined according to the volume data characteristics of each second candidate region. The second matching degree between the second candidate area and the conus medullaris, and determining the second candidate area with the highest second matching degree as the conus medullaris area.

In this embodiment, it is also possible to identify the sagittal plane of the fetus from the three-dimensional volume data of the fetus first, and then identify the conus medullaris region from the sagittal plane image. For example, according to the characteristics of the sagittal plane passing through the fetal spine, the sagittal plane image passing through the fetal spine can be determined from the three-dimensional volume data of the fetus, and then based on the characteristics of the conus medullaris, in the sagittal plane image passing through the fetal spine Identify the area of the conus medullaris. Here, the sagittal plane may be the median sagittal plane and/or a sagittal plane adjacent to the median sagittal plane.

In this embodiment, the lumbar region can also be identified from the three-dimensional volume data of the fetus, and the position of the conus medullaris is displayed relative to the lumbar region, so that the user can easily see the relative positional relationship between the conus medullaris and the lumbar region . For example, a lumbar region may be identified from fetal volumetric data based on characteristics of the fetal lumbar region, an ultrasound image of the lumbar region displayed, and a conus region (e.g., end of the conus, etc.) displayed relative to the ultrasound image of the lumbar region s position. Here, displaying the position of the conus medullaris with respect to the ultrasonic image of the lumbar region may include various suitable methods, for example, the ultrasonic image of the lumbar region and the position of the conus medullaris can be simultaneously displayed, so that the user can directly see the relative position of the conus medullaris. Positional relationship, or display the distance of the conus medullaris region relative to the lumbar region through words or symbols, or display the relative positional relationship between the two through symbols representing the lumbar region and the conus medullaris region, and so on.

In this embodiment, the specific schemes of the various steps involved (for example, identifying the conus medullaris region, identifying the lumbar region, identifying the sagittal plane of the fetus, etc.) can refer to the methods in the foregoing embodiments or with the methods in the foregoing embodiments The similar steps are the same or similar, and will not be described in detail here.

The embodiment of the present application also provides an ultrasonic imaging device 10, as shown in FIG. 1 , including:

Probe 100;

A transmitting circuit 101, which excites the probe 100 to transmit ultrasonic waves to the first tissue to be measured;

The receiving circuit 103 is configured to receive the ultrasonic echo returned from the first tissue to be tested through the probe 100 to obtain an ultrasonic echo signal;

Processor 105, processing the ultrasonic echo signal to obtain a three-dimensional ultrasonic image of the marked fetal rib structure;

A display 106 for displaying a three-dimensional ultrasonic image of the fetal rib structure after marking;

Wherein, the processor 105 also performs the following steps:

Acquiring three-dimensional volume data of the first tissue to be measured according to the ultrasonic echo information; identifying the three-dimensional volume data of the fetal spine and the three-dimensional volume data of the fetal ribs from the three-dimensional volume data of the first tissue to be measured; Rendering the three-dimensional volume data of the fetal spine and the three-dimensional volume data of the fetal ribs to obtain a three-dimensional ultrasonic image of the fetal rib structure; marking the fetal ribs in the three-dimensional ultrasonic image of the fetal rib structure; outputting the marked Three-dimensional ultrasound image of fetal rib structure.

The embodiment of the present application also provides an ultrasonic imaging device 10, as shown in FIG. 1 , including:

Probe 100;

A transmitting circuit 101, which excites the probe 100 to transmit ultrasonic waves to the second tissue to be measured;

The receiving circuit 103 is configured to receive the ultrasonic echo returned from the second tissue to be tested through the probe 100 to obtain an ultrasonic echo signal;

A processor 105, processing the ultrasonic echo signal to obtain a three-dimensional ultrasonic image of the marked vertebral structure;

A display 106 for displaying a three-dimensional ultrasound image of the marked vertebral structure;

Wherein, the processor 105 also performs the following steps:

According to the ultrasonic echo information, obtain the three-dimensional volume data of the second tissue to be measured; identify the three-dimensional volume data of the conus medullaris and the three-dimensional volume data of the lumbar spine from the second three-dimensional volume data; the three-dimensional volume data of the conus medullaris Render the data and the three-dimensional volume data of the lumbar spine to obtain a three-dimensional ultrasonic image of the vertebral structure; mark the conus medullaris in the three-dimensional ultrasonic image of the vertebral structure; output the marked three-dimensional image of the vertebral structure Ultrasound image.

Correspondingly, an embodiment of the present invention further provides a computer-readable storage medium, where a computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, the steps of the above-mentioned ultrasonic imaging method are realized.

The above descriptions of the embodiments of the ultrasonic imaging system and the computer-readable storage medium are similar to the descriptions of the above method embodiments, and have similar beneficial effects to those of the method embodiments. For the technical details not disclosed in the embodiments of the ultrasound imaging system and the computer-readable storage medium of the present invention, please refer to the description of the method embodiment of the present invention for understanding.

In the embodiment of the present application, if the above-mentioned ultrasonic imaging method is implemented in the form of a software function module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiment of the present application is essentially or the part that contributes to the prior art can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for Make a computer device (which may be a personal computer, a server, or a network device, etc.) execute all or part of the methods described in the various embodiments of the present application. The aforementioned storage medium includes: various media that can store program codes such as U disk, mobile hard disk, read-only memory (ROM, Read Only Memory), magnetic disk or optical disk. Thus, embodiments of the present application are not limited to any specific combination of hardware and software.

It should be understood that reference throughout the specification to "one embodiment" or "an embodiment" means that a particular feature, structure, or characteristic related to the embodiment is included in at least one embodiment of the present application. Thus, appearances of "in one embodiment" or "in an embodiment" in various places throughout the specification are not necessarily referring to the same embodiment. Furthermore, the particular features, structures or characteristics may be combined in any suitable manner in one or more embodiments. It should be understood that, in various embodiments of the present application, the sequence numbers of the above-mentioned processes do not mean the order of execution, and the execution order of the processes should be determined by their functions and internal logic, and should not be used in the embodiments of the present application. The implementation process constitutes any limitation. The serial numbers of the above embodiments of the present application are for description only, and do not represent the advantages and disadvantages of the embodiments.

It should be noted that, in this document, the term "comprising", "comprising" or any other variation thereof is intended to cover a non-exclusive inclusion such that a process, method, article or apparatus comprising a set of elements includes not only those elements, It also includes other elements not expressly listed, or elements inherent in the process, method, article, or device. Without further limitations, an element defined by the phrase "comprising a ..." does not preclude the presence of additional identical elements in the process, method, article, or apparatus comprising that element.

In the several embodiments provided in this application, it should be understood that the disclosed devices and methods may be implemented in other ways. The device embodiments described above are only illustrative. For example, the division of the units is only a logical function division. In actual implementation, there may be other division methods, such as: multiple units or components can be combined, or May be integrated into another system, or some features may be ignored, or not implemented. In addition, the coupling, or direct coupling, or communication connection between the components shown or discussed may be through some interfaces, and the indirect coupling or communication connection of devices or units may be electrical, mechanical or other forms of.

The units described above as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units; they may be located in one place or distributed to multiple network units; Part or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

In addition, each functional unit in each embodiment of the present application can be integrated into one processing unit, or each unit can be used as a single unit, or two or more units can be integrated into one unit; the above-mentioned integration The unit can be realized in the form of hardware or in the form of hardware plus software functional unit.

Those of ordinary skill in the art can understand that all or part of the steps to realize the above method embodiments can be completed by hardware related to program instructions, and the aforementioned programs can be stored in computer-readable storage media. When the program is executed, the execution includes The steps of the foregoing method embodiments; and the foregoing storage media include: removable storage devices, read-only memory (Read Only Memory, ROM), magnetic disks or optical disks and other media that can store program codes.

Alternatively, if the above-mentioned integrated units of the present application are realized in the form of software function modules and sold or used as independent products, they can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiment of the present application is essentially or the part that contributes to the prior art can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for Make a computer device (which may be a personal computer, a server, or a network device, etc.) execute all or part of the methods described in the various embodiments of the present application. The aforementioned storage medium includes various media capable of storing program codes such as removable storage devices, ROMs, magnetic disks or optical disks.

The above is only a specific implementation of the application, but the scope of protection of the application is not limited thereto. Anyone familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the application. Should be covered within the protection scope of this application. Therefore, the protection scope of the present application should be determined by the protection scope of the claims.

Claims (12)

1. An ultrasound imaging method, for application to an ultrasound imaging apparatus, the method comprising:

acquiring three-dimensional volume data of a first tissue to be tested;

identifying three-dimensional data of a fetal spinal column and three-dimensional data of fetal ribs from the three-dimensional data of the first tissue to be tested;

rendering the three-dimensional data of the fetal spinal column and the three-dimensional data of the fetal rib to obtain a three-dimensional ultrasonic image of the fetal rib structure; when three-dimensional rendering is carried out on the three-dimensional data of the fetal spinal column and the three-dimensional data of the fetal ribs, three-dimensional ray perspective rendering is carried out on the three-dimensional data of the fetal spinal column and the three-dimensional data of the fetal ribs, so that a three-dimensional ultrasonic image of the fetal rib structure is formed;

marking the fetal ribs in the three-dimensional ultrasonic image of the fetal rib structure and/or determining the number of the fetal ribs according to the identified three-dimensional volume data of the fetal ribs;

Outputting a three-dimensional ultrasonic image of the marked fetal rib structure and/or displaying the determined number of the fetal ribs;

wherein the identifying the three-dimensional data of the fetal spine and the three-dimensional data of the fetal ribs from the three-dimensional data of the first tissue to be tested comprises:

identifying three-dimensional volume data of a fetal rib structure from the three-dimensional volume data of the first tissue to be tested;

dividing the three-dimensional data of the fetal rib structure from the three-dimensional data of the first tissue to be tested;

and identifying the three-dimensional data of the fetal spinal column and the three-dimensional data of the fetal rib from the three-dimensional data of the fetal rib structure.

2. The method of claim 1, wherein identifying three-dimensional volume data of a fetal spine and three-dimensional volume data of a fetal rib from the three-dimensional volume data of the first tissue under test comprises:

based on a first rib detection model, three-dimensional volume data of the fetal spinal column and three-dimensional volume data of the fetal rib are identified from the three-dimensional volume data of the first tissue to be tested.

3. The method of claim 1, wherein the identifying three-dimensional volume data of the fetal rib structure from the three-dimensional volume data of the first tissue under test comprises:

Displaying a three-dimensional image corresponding to the three-dimensional volume data of the first tissue to be tested;

receiving a first input operation based on a corresponding three-dimensional image of the three-dimensional volume data of the first tissue to be tested;

determining a mark point corresponding to the first input operation;

and identifying the three-dimensional volume data of the fetal rib structure from the three-dimensional volume data of the first tissue to be tested according to the coordinates of the mark points.

4. The method of claim 1, wherein the identifying three-dimensional volume data of the fetal rib structure from the three-dimensional volume data of the first tissue under test comprises:

displaying a three-dimensional image corresponding to the three-dimensional volume data of the first tissue to be tested;

receiving a second input operation based on a three-dimensional image corresponding to the three-dimensional volume data of the first tissue to be tested;

determining a first seed region corresponding to the second input operation, wherein the first seed region is positioned in a three-dimensional image region corresponding to the fetal rib structure;

determining a first pixel characteristic of three-dimensional volume data of a first seed region;

and identifying the three-dimensional volume data of the fetal rib structure in the three-dimensional volume data of the first tissue to be tested according to the first pixel characteristics.

5. The method of claim 1, wherein the identifying three-dimensional volume data of the fetal rib structure from the three-dimensional volume data of the first tissue under test comprises:

determining at least two first candidate areas from the three-dimensional volume data of the first tissue to be tested, and acquiring the volume data characteristics of the three-dimensional volume data of each first candidate area;

determining a first matching degree of each first candidate region and the fetal rib structure according to the volume data characteristics of each first candidate region;

determining a first candidate region with highest first matching degree as a target region corresponding to the fetal rib structure;

and taking the three-dimensional data of the target area corresponding to the fetal rib structure as the three-dimensional data of the fetal rib structure.

6. The method of claim 1, wherein the identifying the three-dimensional volume data of the fetal spine and the three-dimensional volume data of the fetal ribs from the three-dimensional volume data of the fetal rib structures comprises:

locating the included spine and fetal ribs from the three-dimensional volume data of the fetal rib structures based on the shape features of the fetal spine and the shape features of the fetal ribs;

Displaying a three-dimensional image corresponding to the three-dimensional volume data of the fetal rib structure;

receiving a third input operation based on a three-dimensional image corresponding to the three-dimensional volume data of the fetal rib structure;

determining a reference fetal rib and a rib identifier corresponding to the reference fetal rib according to the third input operation; the reference fetal rib is at least one fetal rib in the fetal rib structure;

and identifying three-dimensional data of the fetal ribs based on the reference fetal ribs and rib identifiers corresponding to the reference fetal ribs.

7. The method of claim 1, wherein the identifying the three-dimensional volume data of the fetal spine and the three-dimensional volume data of the fetal ribs from the three-dimensional volume data of the fetal rib structures comprises:

based on a second rib detection model, three-dimensional volume data of the fetal spinal column and three-dimensional volume data of the fetal rib are identified from the three-dimensional volume data of the fetal rib structure.

8. The method according to claim 1, wherein the method further comprises:

straightening the three-dimensional data of the fetal rib to obtain straightened rib three-dimensional data;

straightening the three-dimensional data of the fetal spinal column to obtain straightened spinal column three-dimensional data;

Performing plane fitting on the straightened ribs in the straightened rib three-dimensional data and the straightened fetal spine in the straightened spine three-dimensional data to obtain a first plane;

and obtaining image data on the first plane according to the three-dimensional data of the straightened rib and the three-dimensional data of the straightened spine, and obtaining a coronal image of the fetal rib structure.

9. The method according to claim 1, wherein the method further comprises:

determining a target fetal rib from the ribs of the fetal rib structure;

straightening the three-dimensional data of the target fetal rib to obtain straightened three-dimensional data of the target fetal rib;

straightening the three-dimensional data of the fetal spinal column to obtain straightened spinal column three-dimensional data;

taking a plane which is positioned on the same plane with the rib of the straightened target fetus and is perpendicular to the spinal column of the straightened fetus as a second plane;

and obtaining image data on the second plane according to the three-dimensional data of the straightened target fetal rib and the three-dimensional data of the straightened spine, and obtaining a cross-sectional image of the target fetal rib.

10. The method according to claim 8 or 9, wherein straightening the straightened object comprises:

Determining a longitudinal axis of the straightened subject;

sampling the straightened subject according to the longitudinal axis of the straightened subject to obtain a section sequence, wherein the straightened subject comprises fetal ribs or fetal spines;

reconstructing the sequence of slices along a straight line.

11. The method according to claim 1, wherein the method further comprises:

binarizing the three-dimensional data of the fetal ribs and the three-dimensional data of the fetal spinal column to obtain reconstructed three-dimensional data;

rendering the reconstructed three-dimensional data to obtain a three-dimensional skeleton of the fetal rib structure.

12. An ultrasonic imaging apparatus, comprising:

a probe;

a transmitting circuit for exciting the probe to transmit ultrasonic waves to a first tissue to be tested;

a receiving circuit for receiving an ultrasonic echo returned from the first tissue to be tested through the probe to obtain an ultrasonic echo signal;

a processor that processes the ultrasound echo signals to obtain a three-dimensional ultrasound image of the marked fetal rib structure;

a display for displaying a three-dimensional ultrasound image of the marked fetal rib structure;

wherein the processor further performs the steps of:

Acquiring three-dimensional volume data of a first tissue to be tested according to the ultrasonic echo signal;

identifying three-dimensional data of a fetal spinal column and three-dimensional data of fetal ribs from the three-dimensional data of the first tissue to be tested;

rendering the three-dimensional data of the fetal spinal column and the three-dimensional data of the fetal rib to obtain a three-dimensional ultrasonic image of the fetal rib structure; when three-dimensional rendering is carried out on the three-dimensional data of the fetal spinal column and the three-dimensional data of the fetal ribs, three-dimensional ray perspective rendering is carried out on the three-dimensional data of the fetal spinal column and the three-dimensional data of the fetal ribs, so that a three-dimensional ultrasonic image of the fetal rib structure is formed; marking the fetal ribs in the three-dimensional ultrasonic image of the fetal rib structure and/or determining the number of the fetal ribs according to the identified three-dimensional volume data of the fetal ribs;

outputting a three-dimensional ultrasonic image of the marked fetal rib structure and/or displaying the determined number of the fetal ribs;

wherein the identifying the three-dimensional data of the fetal spine and the three-dimensional data of the fetal ribs from the three-dimensional data of the first tissue to be tested comprises:

identifying three-dimensional volume data of a fetal rib structure from the three-dimensional volume data of the first tissue to be tested;

Dividing the three-dimensional data of the fetal rib structure from the three-dimensional data of the first tissue to be tested;

and identifying the three-dimensional data of the fetal spinal column and the three-dimensional data of the fetal rib from the three-dimensional data of the fetal rib structure.

Images