CN103300856A - Positioning method and positioning device for cervical vertebral body axial lines and relevant tissues of MRI (magnetic resonance imaging) images - Google Patents

Abstract

The invention discloses a method and a device for positioning the axis of a cervical vertebral body and related tissues. The present invention quickly locates the axis of the cervical spine by locating the trachea line, and utilizes the characteristics that the gray scale in the vertebral body is approximately uniform, has a clear boundary with the intervertebral disc, and the shape and size of the vertebral body are basically fixed to adaptively extract the vertebral body. The extraction process is not affected. Based on the influence of image weight, the midline position and angle of each intervertebral disc can be calculated according to the extracted vertebral body, so as to effectively automatically position the cervical intervertebral disc in the MRI system. The invention also discloses a magnetic resonance system at the same time.

Description

Method and device for locating cervical vertebral body axis and related tissues in MRI image

technical field

The invention relates to medical imaging equipment, in particular to a method and device for automatic positioning of cervical intervertebral discs by using magnetic resonance images, and a magnetic resonance imaging system.

Background technique

Magnetic Resonance Imaging (MRI, Magnetic Resonance Imaging) has been widely used clinically due to the advantages of non-invasive and multi-parameter imaging, especially its ability of arbitrary tomographic imaging, which can directly observe and analyze tissue structure and its lesions from different angles , so it has outstanding advantages in the serial examination of the spine. A typical MRI examination process of cervical intervertebral disc often requires physicians to manually place each group of scanning lines on the intervertebral disc with disease on the sagittal plane positioning image. In order to ensure that the thread group passes through the center of the intervertebral disc, it is necessary to repeatedly adjust the position and angle of the thread group. This process is complicated and time-consuming. Therefore, if the automatic identification of MRI intervertebral discs can be realized, the intelligent scanning and positioning of the intervertebral discs can be realized, thereby reducing the scanning time and the operating burden of physicians. The existing automatic intervertebral disc extraction technology mainly adopts an image processing method to automatically segment or identify a vertebral body or an intervertebral disc in a sagittal plane image. Some traditional image segmentation processing methods mainly find the region to be segmented through edge detection or prior shape information, and then realize the segmentation according to the gray value range of the target region or the difference with the neighborhood. However, this segmentation method often takes a long time and is not very effective, because in the MRI images with different weights, the gray value of the intervertebral disc will change. Recovery, short T1 inversion recovery) weight image is white. In addition, there are some methods that use deformable model matching to extract the vertebral body and then locate the intervertebral disc. The problem with this method is that the number of visible vertebral bodies in the spine image is not necessarily the same. If there are too many or too few visible vertebral bodies, the matching will fail. . Other segmentation methods require physicians to perform certain interactive operations, such as selecting feature points, which will reduce the efficiency of inspection.

Contents of the invention

The main technical problem to be solved by the present invention is to provide a method and device for automatic positioning of cervical intervertebral discs using magnetic resonance images, and a magnetic resonance imaging system.

According to one aspect of the present invention, a method and device for positioning the axis of the cervical vertebral body are provided, wherein the device includes: a tracheal line positioning unit, which is used to detect magnetic The body surface boundary image of the resonance image is obtained to obtain a body surface boundary image, and the trachea line is detected according to the body surface boundary image; the vertebral body axis positioning unit is used to align the airway line according to the preset translation conditions according to the position characteristics of the trachea line and the cervical vertebral body axis. The coordinates of the pipeline are translated to obtain the coordinates of the cervical vertebral body axis, so as to locate the vertebral body axis.

According to another aspect of the present invention, a cervical vertebral body positioning method and device are provided, wherein the device includes: the above-mentioned positioning device for the cervical vertebral body axis; degree change gradient to obtain the possible boundary points of the vertebral body, filter the possible boundary points according to the third preset condition, move the coordinates of the filtered boundary points along the direction of the vertebral body axis, and the corresponding point of the obtained new coordinates is The seed point inside the vertebral body; the vertebral body positioning unit is used to obtain the most probable vertebral body region by using the region growing method based on the seed point inside the vertebral body.

According to another aspect of the present invention, a cervical vertebral disc positioning method and device are provided, wherein the device includes: the cervical vertebral body positioning device as described above; The scanning method detects the vertices of the vertebral body area; the intervertebral disc centerline determination unit is used to obtain two center points according to the connection line between the vertices of two adjacent vertebral body regions, and the connection line of the two center points is the intervertebral disc center line.

The present invention also provides a magnetic resonance imaging system comprising the above-mentioned cervical vertebral body positioning device or cervical intervertebral disc positioning device.

Description of drawings

Fig. 1 is a schematic structural diagram of a magnetic resonance imaging system in an embodiment of the present invention;

Fig. 2 is a schematic structural view of a cervical vertebral body positioning device in an embodiment of the present invention;

Fig. 3 is a schematic flow chart of a cervical vertebral body positioning method in an embodiment of the present invention;

Figure 4 is a T1 weighted sagittal cervical spine image;

Fig. 5 is a schematic flow chart of positioning the axis of the vertebral body in an embodiment of the present invention;

Fig. 6 is a schematic diagram of a set of maximum/minimum points detected after obtaining the first derivative;

Fig. 7 is a schematic diagram of a body surface boundary detected in an embodiment of the present invention;

Fig. 8 is a schematic diagram of the extraction results of extreme points after seeking the second derivative in an embodiment of the present invention;

Fig. 9 is a schematic diagram of the air pipeline obtained after screening in an embodiment of the present invention;

Fig. 10 is a schematic diagram of the axis of the vertebral body obtained in an embodiment of the present invention;

Fig. 11 is a schematic flow diagram of selecting a seed point inside a vertebral body in an embodiment of the present invention;

Fig. 12 is a schematic flow chart of extracting the vertebral body region in an embodiment of the present invention;

Fig. 13 is a schematic diagram of the vertebral body region located in an embodiment of the present invention;

Fig. 14 is a schematic flowchart of a cervical intervertebral disc positioning method according to an embodiment of the present invention;

Fig. 15 is a schematic diagram of a scanning path for detecting vertebral body corners in an embodiment of the present invention;

Fig. 16 is a schematic diagram of the centerline of the intervertebral disc extracted in an embodiment of the present invention.

Detailed ways

The present invention will be further described in detail below through specific embodiments in conjunction with the accompanying drawings.

In each embodiment of the present invention, the identification and positioning of the cervical vertebral body and the cervical intervertebral disc usually involves first positioning the "vertebral body axis" and the "trachea line". concept. The said vertebral body axis in each embodiment of the present invention is not strictly limited to the line passing through the center of the vertebral body, as long as it can pass through all the vertebral bodies longitudinally, so the vertebral body axis defined in the various embodiments of the present invention is Some errors are allowed. Since the trachea region is in the shape of a long and narrow band in the entire magnetic resonance image, the detected trachea region is in the same line, so the trachea region is described here as a trachea line.

Fig. 1 shows the structure of a magnetic resonance imaging system in an embodiment of the present invention. As shown in FIG. 1 , the magnetic resonance imaging system 100 includes a magnet system 110 , a gradient magnetic field system 120 , a radio frequency system 130 and a control and processing system 140 . The magnet system 110 includes a magnet 111, a gradient magnetic field coil 112, a transmitting coil 113 and a receiving coil 114. The magnet 111 can be a permanent magnet or a constant conduction magnet, which is used to provide a constant main magnetic field to the object to be measured (such as a patient). The coil 112 is used to generate a gradient magnetic field in three-dimensional space, the transmitting coil 113 is used to provide radio frequency (RF) pulses to excite the spins of atomic nuclei in the object to be tested, and the receiving coil 114 is used to detect echo signals emitted by the object to be tested. The gradient magnetic field system 120 is connected to the control and processing system 140 for driving the gradient magnetic field coil 112 under the control of the control and processing system 140 . The radio frequency system 130 is connected to the control and processing system 140 for generating RF pulses under the control of the control and processing system 140 and applying them to the transmitting coil 113 after being amplified. The control and processing system 140 is used not only for controlling various parts, but also for processing echo signals. The echo signal detected by the receiving coil 114 is transmitted to the control and processing system 140 .

In one embodiment, the control and processing system 140 includes a cervical vertebral body positioning device, configured to locate the cervical vertebral body of the subject on the magnetic resonance image based on the obtained magnetic resonance image. In another embodiment, the control and processing system 140 includes a cervical intervertebral disc positioning device, which is used to identify the centerline of the cervical intervertebral disc of the subject on the magnetic resonance image based on the obtained magnetic resonance image, thereby reducing the doctor's operation burden.

Fig. 2 shows the structure of a cervical vertebral body positioning device in an embodiment. As shown in FIG. 2 , the cervical vertebral body positioning device 200 includes a cervical vertebral body axis positioning device 210 , a seed point selection unit 230 and a vertebral body positioning unit 250 . The positioning device 210 of the cervical vertebral body axis is used to determine the vertebral body axis according to the trachea line; the seed point selection unit 230 is used to obtain the possible boundary points of the vertebral body according to the gray scale gradient of the vertebral body axis, according to the third preset condition Screen the possible boundary points, move the coordinates of the screened boundary points along the direction of the vertebral body axis, and the point corresponding to the obtained new coordinates is the seed point inside the vertebral body; the vertebral body positioning unit 250 is used to For the internal seed point, the most probable vertebral body region is obtained by region growing method.

Still as shown in Figure 2, in a specific example, the positioning device 210 of the axis of the cervical vertebral body includes a tracheal line positioning unit 211 and an axis positioning unit 213; wherein, the tracheal line positioning unit 211 is used to utilize the measured According to the transition characteristics between the body of the patient and the imaging background, the body surface boundary of the magnetic resonance image is detected to obtain the body surface boundary, and the trachea line is detected according to the body surface boundary; Position characteristics, the coordinates of the trachea line are translated according to the preset translation conditions to obtain the coordinates of the cervical vertebral body axis, so as to locate the vertebral body axis.

In another specific example, still as shown in FIG. 2 , the vertebral body positioning unit 250 includes an initial calculation subunit 251 , a cycle calculation subunit 253 and a growth subunit 255 . Among them, the initial calculation subunit 251 is used to initialize the growth threshold, perform region growth based on the initialized growth threshold, fill the grown region, calculate the area-perimeter ratio of the filled region, and estimate the expected The area-to-perimeter ratio of the area to obtain the difference between the area-to-perimeter ratio of the filled region and the expected area-to-perimeter ratio; the loop calculation subunit 253 is used to change the growth threshold, and re-grow the region according to the changed growth threshold, and will The grown area is filled to obtain the new area-to-perimeter ratio of the filled area, the new estimated expected area-to-perimeter ratio, and the difference between the two, and this loops until the difference is the smallest and the filling The area of the rear region is smaller than the maximum area of the preset growth target, and the area of the filled region is larger than the minimum area of the preset growth target; the growth subunit 255 is used to regard the growth threshold corresponding to the minimum difference as the optimal growth threshold, and according to the growth threshold And grow based on the seed point inside the vertebral body to get the preliminary vertebral body area. In yet another specific example, besides the initial calculation subunit 251, the cycle calculation subunit 253 and the growth subunit 255, the vertebral body positioning unit 250 also includes an optimization subunit 257, which is used to obtain the preliminary vertebral body area, According to the region information of each seed point after region growth, the optimal vertebral body region is obtained; wherein, the region information includes the position information of the region after the growth of the adjacent seed point, and the area difference of the region after the growth of the adjacent seed point.

Based on the above device, a cervical vertebral body positioning method is shown in Figure 3, comprising the following steps:

Step S1, locating the axis of the cervical vertebral body;

Step S2, selecting seed points based on the axis of the vertebral body;

Step S3, acquiring a vertebral body based on the seed point.

The above steps will be described in detail below with reference to FIG. 4 to FIG. 13 through embodiments.

For step S1, the cervical vertebral body axis is located by locating the trachea line by using the characteristics of the trachea region in the magnetic resonance image.

In the MRI images of the cervical spine, the trachea area is a relatively significant area, because there is no tissue such as water or fat in the trachea, it appears black in the MRI images of different weights, which is obviously different from the surrounding tissues, as shown in Figure 4 , the tracheal region is darker than the surrounding tissue. In FIG. 4 , the left side of the figure is the front of the subject, and the right side of the figure is the back of the subject.

Since the cervical spine is next to the trachea and its direction is basically the same, in one embodiment, the trachea is firstly positioned to locate the cervical spine area. The process of locating the trachea can be subdivided into: extracting the body surface boundary on the first side (for example, the left side) of the measured object, extracting the body surface boundary on the second side (for example, the right side), and locating the trachea. Therefore, in this embodiment, as shown in FIG. 5, step S1 includes the following steps:

Step S11, detecting body surface boundaries. The extraction of the body surface boundary can utilize the transition characteristics between the body and the imaging background. For example, as shown in FIG. 4 , the body surface on the left is a transition from black to white, and the body surface on the right is a transition from white to black. This transition can be detected by using the first derivative, and since the body surface boundary is roughly vertical, it can be detected by the extreme value of the first horizontal derivative. Calculate the first-order derivative in the direction perpendicular to the body surface boundary direction (such as the horizontal direction), and the first-order derivative can be calculated according to the following expression:

$\frac{<span\; class="notranslate">\; \&PartialD;</span>\mathrm{<span\; class="notranslate">\; I</span>}}{<span\; class="notranslate">\; \&PartialD;</span>\mathrm{<span\; class="notranslate">\; x</span>}}<span\; class="notranslate">\; *</span>\mathrm{<span\; class="notranslate">\; G</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; I</span>}<span\; class="notranslate">\; *</span>\frac{<span\; class="notranslate">\; \&PartialD;</span>\mathrm{<span\; class="notranslate">\; G</span>}}{<span\; class="notranslate">\; \&PartialD;</span>\mathrm{<span\; class="notranslate">\; x</span>}}$

表示x方向(水平方向)一阶导数，G为高斯模板，*为卷积。Among them, I is the input image,

Indicates the first derivative in the x direction (horizontal direction), G is the Gaussian template, and * is the convolution.

After selecting an appropriate Gaussian template to calculate the first-order derivative, detect the maximum and minimum points of the first-order derivative in the horizontal direction, respectively. The maximum point is used to detect the left body surface boundary, and the minimum point is used for Detect the right body surface boundary. Transform the position of the first derivative along the direction of the body surface boundary, repeat the above steps, and detect many maximum and minimum points. Some maximum points are connected into lines to form a set of maximum points, and some minimum points The points are connected into a line to form a set of minimum points. The maximum value/minimum value point image is shown in Figure 6, where the white point is the maximum value and the gray point is the minimum value. It can be seen that there are too many maximum value point sets and minimum value point sets in the image, therefore, it is necessary to filter multiple maximum value point sets and minimum value point sets according to the first preset condition, the first The preset condition may be comprehensively considered according to the length of the line, the derivative value of each point on the line, and the distance between the line and the image boundary. The body surface boundary is obtained after screening, as shown in FIG. 7 , only the first side (for example, left side) body surface boundary and the second side (for example, right side) body surface boundary are left after screening by the first preset condition. Step S12 is executed after the body surface boundary is obtained. Certainly, those skilled in the art should understand that the first preset condition may be other screening conditions besides the screening conditions disclosed in this embodiment, as long as the body surface boundaries on both sides are screened out from many boundaries.

Step S12, detecting the gas pipeline. After extracting the body surface boundary, use the left and right range defined by the body surface boundary to detect the trachea line. The trachea line is in the shape of a long and narrow strip in the image, and its gray value is close to 0 (black), which is in sharp contrast with its left and right sides. Using this feature, the second derivative can be used to detect gas pipelines. The second derivative is sensitive to linear structures, but needs to specify an appropriate filter scale, and can only be effectively detected when the filter scale matches the width of the gas line. Because the body surface boundary has been extracted, the distance between the first side body surface boundary and the second side body surface boundary can be used to estimate the filter scale, and the estimated filter scale is used to calculate the second derivative of the magnetic resonance image; the second derivative calculation expression as follows:

$\mathrm{<span\; class="notranslate">\; I</span>}<span\; class="notranslate">\; *</span>\frac{{<span\; class="notranslate">\; \&PartialD;</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}\mathrm{<span\; class="notranslate">\; G</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; \&sigma;</span>}<span\; class="notranslate">\; )</span>}{{<span\; class="notranslate">\; \&PartialD;</span>\mathrm{<span\; class="notranslate">\; x</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}<span\; class="notranslate">\; ,</span>$ $\mathrm{<span\; class="notranslate">\; \&sigma;</span>}<span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; d</span>}}_{\mathrm{<span\; class="notranslate">\; lr</span>}}}{\mathrm{<span\; class="notranslate">\; m</span>}}$

Among them, σ is the standard deviation of the Gaussian template, which is estimated according to the distance between the body surface boundary on the first side and the body surface boundary on the second side, reflecting the filtering scale; d _lr is the average distance between the body surface boundaries on both sides, and M is a proportional constant, which can be Determined by experiment.

After obtaining the second-order derivative and filtering the image to obtain the extreme value, that is, to detect the extreme value of the second-order derivative image, so as to obtain a plurality of extreme point sets connected by the extreme point points into a line, the extraction result of the extreme point is shown in Figure 8. After the extreme points are extracted, screening still needs to be performed. At this time, the second preset condition (i.e., the screening rule) can be comprehensively considered according to factors such as the horizontal position, vertical position, and length of the trachea line in the cervical spine image. For example, this In the embodiment, according to the image obtained from the side of the subject, as shown in Figure 4, the horizontal position of the air line is on the left, the vertical position is on the bottom, and then the length of the air line and other factors are considered comprehensively, so as to obtain the air line after screening, As shown in Figure 9. In addition to the conditions disclosed in this embodiment, the second preset condition may also be other screening conditions, as long as the gas pipelines are screened out from many boundaries.

Step S13, locating the axis of the vertebral body. Since the axis of the cervical vertebral body is close to the right side of the trachea and approximates the same direction as the trachea, it can be obtained by translating the trachea line. In addition, considering that the closer to the top of the trachea, the closer it is to the cervical spine, therefore, a simple correction can be made to the translation parameter: that is, the coordinate translation of the pixel point at the uppermost end of the trachea line is the smallest (set as the first translation parameter), and the most The coordinate translation of the pixel point at the lower end is the largest (set as the second translation parameter), and the translation parameter of the coordinates of the pixel point in the middle can be obtained by interpolation (such as linear interpolation), and the first translation parameter and the second translation parameter can also use two The distance between the lateral body surface boundary is estimated, that is, the first translation parameter and the second translation parameter are empirical values related to the distance between the first lateral body surface boundary and the second lateral body surface boundary. Of course, as mentioned above, the axis of the vertebral body involved in the embodiment is the basis for subsequent positioning of the vertebral body, so the axis is not strictly limited to a line passing through the center of the vertebral body, as long as it can pass through all the vertebral bodies longitudinally, So a certain error is allowed. In one embodiment, the calculation expression of the axis of the vertebral body is as follows:

spine(t)=I(x(t)+k(t)·d _lr , y(t))

Among them, spine(t) represents the axis of the vertebral body, t is the index, (x(t), y(t)) is the coordinate of the trachea line, and k(t) is the proportional coefficient of the translation distance relative to the distance between the left and right body surface boundaries. After calculation, the position of the axis of the vertebral body is shown by the white line on the right in Figure 10.

After the axis of the vertebral body is determined, step S2 is used to select the seed point inside the vertebral body, and step S3 is used to perform region growth based on the seed point, so that the vertebral body region can be obtained.

For step S2, a seed point inside the vertebral body is selected based on the located vertebral body axis. The method of selecting the seed point inside the vertebral body in the embodiment mainly utilizes the two characteristics of a clear boundary between the vertebral body and the intervertebral disc, and a certain distance between the vertebral bodies. The process of selecting seed points can be divided into: calculating the boundary points of the vertebral body, screening the boundary points, and moving the boundary points. As shown in Figure 11, step S2 includes the following steps:

Step S21, determining possible boundary points. Since the gray value at the boundary point of the vertebral body changes significantly, the boundary point can be detected based on the gradient feature. Using the axis of the vertebral body located in step S1, after calculating the first derivative, the position where the gray level changes drastically corresponds to the boundary between the vertebral body and the intervertebral disc. A specific judgment can use |g(t)|>T as a screening condition, and the points that meet this condition are regarded as possible boundary points, where g(t) is the axis gradient, T is the threshold, and T=mean(| g(t)|). Because what adopt in this embodiment is the change gradient of the gray scale on the axis of the vertebral body, the gradient represents the change of the gray scale, it has nothing to do with the height of the gray scale itself, as long as the gray scale changes, it can be reflected on the gradient; so this paper In the embodiment, the seed point inside the vertebral body is not affected by the inconsistency of the gray scale of the vertebral body or the intervertebral disc, and the seed point inside the vertebral body can be extracted more accurately for subsequent growth of the vertebral body region.

Step S22, screening possible boundary points.

The possible boundary points detected in step S21 may be redundant. For example, a boundary often detects multiple boundary points, which requires boundary point screening. Filter according to the third preset condition, such as using the characteristic that there is a certain distance between two adjacent boundary points, so that the distance d between any two adjacent boundary points satisfies:

$\frac{{\mathrm{<span\; class="notranslate">\; d</span>}}_{\mathrm{<span\; class="notranslate">\; lr</span>}}}{{\mathrm{<span\; class="notranslate">\; <figure-callout\; id="1"\; label="N"\; filenames="HDA0000142970190000062.png"\; state="\{\{state\}\}">N</figure-callout></span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}}<span\; class="notranslate">\; \&le;</span>\mathrm{<span\; class="notranslate">\; d</span>}<span\; class="notranslate">\; \&le;</span>\frac{{\mathrm{<span\; class="notranslate">\; d</span>}}_{\mathrm{<span\; class="notranslate">\; lr</span>}}}{{\mathrm{<span\; class="notranslate">\; <figure-callout\; id="2"\; label="N"\; filenames="HDA0000142970190000062.png"\; state="\{\{state\}\}">N</figure-callout></span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}}<span\; class="notranslate">\; ,</span>$

Wherein, N ₁ and N ₂ are empirical values, which can be set according to experience.

In addition to the above-mentioned screening conditions in this embodiment, the third preset condition may also be other screening conditions, as long as the qualified boundary points are selected from many boundary points, so as to reduce the number of boundary points.

Step S23, acquiring a seed point, that is, moving the coordinates of the screened boundary points along the axis of the vertebral body, and the point corresponding to the obtained new coordinates is the seed point inside the vertebral body. For example, the screened boundary points are moved up or down vertically, with the purpose of moving the position of the boundary points to the inside of the vertebral body. The moving distance can be set according to experience, for example, a value greater than the upper and lower width of the intervertebral disc.

For step S3, based on the seed points inside the vertebral body selected in step S2, adaptive threshold region growing is performed to obtain the vertebral body region. In this step, first initialize the growth threshold, perform region growth based on the initialized growth threshold, fill the grown region, calculate the area-to-perimeter ratio of the filled region, and estimate the desired area and circumference based on the filled region length ratio, to obtain the difference between the area-perimeter ratio of the filled region and the expected area-perimeter ratio; then, change the growth threshold, and re-grow the region according to the changed growth threshold, and fill the grown region, Get the area-to-perimeter ratio of the new filled area, the new estimated expected area-to-perimeter ratio, and the difference between the two, and cycle until the difference between the two is the smallest, and the area of the filled area is smaller than the preset The maximum area of the growth target, and the area of the filled area is greater than the minimum area of the preset growth target; finally, the growth threshold corresponding to the minimum difference is regarded as the optimal growth threshold, and the growth is performed according to the growth threshold and based on the seed point inside the vertebral body, and the obtained Preliminary vertebral body area.

Here, the area of the vertebral body and the shape of the vertebral body are used to iteratively calculate the growth threshold. The area coordinates after growth satisfy:

$\underset{\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; K</span>}}{\mathrm{<span\; class="notranslate">\; \&cup;</span>}}}<span\; class="notranslate">\; \{</span><span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>\mathrm{<span\; class="notranslate">\; I</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; I</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; seedx</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; seedy</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; |</span><span\; class="notranslate">\; <</span>{\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; \}</span>$

Among them, (seedx _j , seedy _j ) is the coordinates of the jth seed, ( _xi , y _i ) is the isolated neighborhood set of (seedx _j , seedy _j ), K is the number of seed points H _j is based on the jth The growth threshold of the seed. The key here is to determine the optimal growth threshold H _j .

As shown in Figure 12, taking a single seed as an example, the region growing method to extract the vertebral body region includes the following steps:

和

L₁和L₂依经验选择。Step S31, setting the initial threshold H _ini , the minimum area SM of the growth target, and the maximum area ST of the growth target, and assigning the initial threshold H _ini to the growth threshold H; where the initial threshold H _ini can be a relatively large fixed value, or a vertical A value obtained by multiplying the difference between the maximum gray level and the minimum gray level of the body axis by a coefficient, SM and ST can be estimated by using d _lr , such as

and

L ₁ and L ₂ are selected empirically.

Step S32 , perform region growth based on the growth threshold, and perform region filling on the grown region. In this step, the region growing can use the existing technology to perform region growing and filling based on the seed point.

Step S33, calculating the area perimeter ratio SP of the filled area, Among them, S is the area of the filled area, and P is the perimeter of the filled area.

Step S34, calculating the desired area-perimeter ratio according to the area of the filled area. According to the feature that the shape of the vertebral body is basically fixed and approximately square, assuming that the shape of the vertebral body is a square, the perimeter is calculated according to the area S of the filled area, and then the area-to-perimeter ratio is calculated again, that is, the expected area-to-perimeter ratio is calculated SPR, $\mathrm{SPR}=\frac{S}{4\sqrt{S}-4}.$

Step S35, calculating the difference between the area-perimeter ratio and the expected area-perimeter ratio. In this step, the difference between the area-to-perimeter ratio and the expected area-to-perimeter ratio may be a difference between the two, or a ratio between the two.

Step S36, judging whether the growth satisfies a predetermined condition. The predetermined condition may be that the growth threshold H is already very small and is smaller than the set threshold. The predetermined condition may also be that the area S of the filled region is already very small and smaller than a set threshold. If the growth satisfies the predetermined condition, then step S38 is executed to stop the growth based on the seed, and find the minimum difference in the difference between the area after filling between the minimum area of the growth target and the maximum area of the growth target, and find the minimum difference by the minimum difference. The corresponding growth threshold is recorded as the optimal growth threshold, and the filled area corresponding to the minimum difference is recorded as the cone area based on the seed growth; otherwise, step S47 is performed;

Step S47, transforming the growth threshold. For example, the growth threshold is reduced according to a predetermined rule, and then the process turns to step S42, where the region is grown based on the new growth threshold, and is executed in a loop until the growth meets the predetermined condition. The predetermined rule may be to gradually decrease the growth threshold according to a set step size, or to gradually decrease the growth threshold according to a set curve, or to gradually decrease the growth threshold irregularly or randomly.

Because the area of the vertebral body cannot be accurately estimated, if only the area of the vertebral body is used to determine the threshold, it is likely to obtain an irregularly shaped area, and the segmentation effect is not good. Due to the influence of noise and other factors, the uniformity of the gray level of the vertebral body is not ideal, and the gray level distribution is irregular. The areas grown by different gray level thresholds may have a small difference in area, but the shape is significantly different. Only the extracted area and The shape of the vertebrae is the most regular when they are just matched. Therefore, this embodiment combines the shape information of the vertebral body and the area of the vertebral body, so that the grown area is most similar to the real vertebral body area.

For each seed point, after obtaining the optimal H, the vertebral body region corresponding to the seed point is also obtained through region growth, which is the preliminary vertebral body region.

In another embodiment, the preliminary vertebral body area obtained in step S3 is also analyzed and judged, and a better vertebral body area is obtained through optimization. This is because there is still redundancy in the number of seed points obtained earlier. There are mainly two Cases:

1) Some seed points are not inside the vertebral body, but on the intervertebral disc. At this time, even if the above-mentioned regional growth is performed, the obtained area is not the vertebral body area;

2) Some seed points are in the same cone at the same time, and the areas they grow will overlap.

In order to solve the above two problems, this embodiment analyzes the obtained preliminary vertebral body area, and obtains the best vertebral body area according to the area information of each seed point after area growth; wherein, the area information includes each seed point The difference between SP and SPR after point growth, the position information of the area after the growth of the two seed points before and after, the area difference of the area after the growth of the two seed points before and after, etc., these information can effectively solve the above two problems.

After region growth and region analysis and judgment, the obtained vertebral body region is shown in Figure 13.

According to the content disclosed in the present invention, those skilled in the art should understand that when performing cervical vertebral body extraction, all the steps in the above embodiments can be used, and some steps can also be combined. If there is technology, the scheme in the embodiment of the present invention is adopted when locating the internal seed point of the vertebral body and performing regional growth, or the scheme in the embodiment of the present invention is adopted when positioning the trachea line and performing regional growth, while other steps adopt the existing technology.

The above is the description of the cervical vertebral body positioning device and the corresponding method. In MRI images, the centerline of the cervical intervertebral disc of the subject is more often used to identify the centerline of the cervical intervertebral disc of the subject, so as to reduce the scanning time and reduce the operating burden of the physician. Therefore, a control and processing system in one embodiment includes a cervical intervertebral disc positioning device, which includes the cervical vertebral body positioning device in any of the above embodiments, a corner point determination unit, and an intervertebral disc centerline determination unit. Among them, the corner point determination unit is used to detect the apex of the vertebral body area by using the helical scanning method according to the located cervical vertebral body area; , two center points are obtained, and the connecting line between the two center points is the center line of the intervertebral disc.

Based on the above-mentioned cervical intervertebral disc positioning device, a cervical intervertebral disc positioning method is shown in Figure 14, including the following steps S1-S5:

Step S1, locating the axis of the cervical vertebral body;

Step S2, selecting a seed point based on the axis of the vertebral body;

Step S3, obtaining the vertebral body based on the seed point;

Step S4, determining the corner point of the vertebral body based on the vertebral body;

Since the shape of the vertebral body is basically fixed and approximately square, based on the result of the vertebral body area obtained earlier, it is necessary to detect four vertex coordinates of each vertebral body, namely, the vertex coordinates of the upper left corner, the vertex coordinates of the lower left corner, and the vertex coordinates of the upper right corner , the coordinates of the lower right corner vertex. The detection method may use a helical scanning method to extract the apex of the vertebral body. Take the detection of the upper left corner vertex of a certain vertebral body as an example (set the positive direction of the Cartesian coordinates to be rightward and downward), through the circumscribed rectangle of the vertebral body, scan the circumscribed rectangle spirally at a predetermined angle such as 45 degrees , the coordinates of the upper left corner vertex of the vertebral body area are at the leftmost in the horizontal direction and the uppermost in the vertical direction within the rectangular frame. The scan path is shown in Figure 15. Similarly, the four vertices of the vertebral body can be obtained respectively by adopting different scanning directions.

Step S5, determining the centerline of the intervertebral disc based on the corner points of the vertebral body.

The embodiment uses the vertices of two adjacent vertebral bodies to calculate the center point of the intervertebral disc. For example, the midpoint coordinates of the apex of the lower left corner of the previous vertebral body and the apex of the upper left corner of the posterior vertebral body are used as the left center point of the intervertebral disc, and the midpoint coordinates of the apex of the lower right corner of the previous vertebral body and the apex of the upper right corner of the posterior vertebral body are used as the intervertebral disc , the centerline of the intervertebral disc can be determined through these two center points. The final extracted centerline of the intervertebral disc is shown in Figure 16.

Steps S1 to S3 in this embodiment may be implemented by using all or part of the steps provided by the present invention, or all or part of them may be implemented by using existing technologies.

The present invention quickly locates the cervical spine by locating the trachea, and utilizes the characteristics that the gray scale in the vertebral body is approximately uniform, has a clear boundary with the intervertebral disc, and the shape and size of the vertebral body are basically fixed to adaptively extract the vertebral body without being affected by the weight of the image , and then calculate the midline position and angle of each intervertebral disc, which can be effectively applied to the automatic positioning of cervical intervertebral disc scanning in the MRI system.

The invention is not affected by high or low field strength, and can be applied to MRI imaging systems of any field strength.

The above describes how to locate the cervical vertebral body or the cervical intervertebral disc based on the magnetic resonance image, but when other methods are used to obtain the tissue image of the object under test, the method and/or device provided by the present invention can also be used to locate the cervical vertebral body or the cervical intervertebral disc.

The above content is a further detailed description of the present invention in conjunction with specific embodiments, and it cannot be assumed that the specific implementation of the present invention is limited to these descriptions. For those of ordinary skill in the technical field of the present invention, without departing from the concept of the present invention, some simple deduction or replacement can be made, which should be regarded as belonging to the protection scope of the present invention.

Claims (10)

1. a positioning method of cervical vertebral axis, it is characterized in that comprising: The trachea line positioning step is to detect the body surface boundary of the magnetic resonance image by using the transition characteristics of the subject's body and the imaging background in the magnetic resonance image, and detect the trachea line according to the body surface boundary; In the step of locating the vertebral body axis, according to the position characteristics of the tracheal line and the cervical vertebral body axis, the coordinates of the tracheal line are translated according to the preset translation conditions to obtain the coordinates of the cervical vertebral body axis, so as to locate the vertebral body axis. 2. the positioning method of cervical vertebral body axis as claimed in claim 1, is characterized in that, in described trachea line positioning step, The body surface boundary of the detected magnetic resonance image includes: The first-order calculation sub-step is to calculate the first-order derivative in the direction perpendicular to the direction of the body surface boundary according to the trend characteristic of the body surface boundary, and obtain the maximum value point set and the minimum value point set of the first-order derivative according to the calculation result, The maximum value point set is used to detect the first side body surface boundary, and the minimum value point set is used to detect the second side body surface boundary; The first screening sub-step is to filter the set of maximum value points and the set of minimum value points according to the first preset condition to obtain the boundary of the body surface of the first side and the boundary of the body surface of the second side; Said detection of the gas line according to said body surface boundary includes: The estimation sub-step is to estimate the filter scale according to the distance between the body surface boundary on the first side and the body surface boundary on the second side; The second-order calculation sub-step is to calculate the second-order derivative of the magnetic resonance image according to the filter scale according to the grayscale characteristics of the trachea in the magnetic resonance image, and obtain the extreme point set of the second-order derivative according to the calculation result, press The second preset condition filters the set of extreme points to obtain the gas pipeline. 3. The positioning method of the cervical vertebral body axis according to claim 1 or 2, wherein the preset translation conditions include: the translation displacement of the pixel point coordinates at the uppermost end of the trachea line is the first translation parameter, and the uppermost end of the trachea line is the first translation parameter. The translation displacement of the pixel point coordinates at the lower end is the second translation parameter, the first translation parameter and the second translation parameter are empirical values related to the distance between the first side body surface boundary and the second side body surface boundary, and the first The translation parameter is smaller than the second translation parameter, and the translation displacement of the pixel point coordinates between the uppermost end and the lowermost end of the air pipeline is obtained by interpolating the first translation parameter and the second translation parameter. 4. A positioning device for cervical vertebral body axis, characterized in that it comprises: The gas line positioning unit is used to detect the body surface boundary of the magnetic resonance image by using the transition characteristics of the subject's body and the imaging background in the magnetic resonance image, and detect the gas line according to the body surface boundary; The vertebral body axis positioning unit is used to translate the coordinates of the tracheal line according to the preset translation conditions to obtain the coordinates of the cervical vertebral body axis according to the position characteristics of the tracheal line and the cervical vertebral body axis, so as to locate the vertebral body axis. 5. A cervical vertebral body positioning method, characterized in that it comprises: The positioning method of the cervical vertebral axis as claimed in any one of claims 1-3; The seed point selection step is to obtain the possible boundary points of the vertebral body according to the gradient of the gray scale change of the vertebral body axis, filter the possible boundary points according to the third preset condition, and set the coordinates of the filtered boundary points along the axis of the vertebral body direction, the point corresponding to the obtained new coordinates is the seed point inside the vertebral body; In the step of acquiring the vertebral body, based on the seed points inside the vertebral body, the preliminary vertebral body region is obtained by using the region growing method and combining the shape and area information of the vertebral body. 6. cervical vertebral body positioning method as claimed in claim 5, is characterized in that, described employing region growth method and combining the shape and area information of vertebral body to obtain preliminary vertebral body region comprises: The initial calculation sub-step, initialize the growth threshold, perform region growth based on the initialized growth threshold, fill the grown region, calculate the area-to-perimeter ratio of the filled region, and estimate the expected area and perimeter based on the filled region Ratio, get the difference between the area-perimeter ratio of the filled area and the expected area-perimeter ratio; Cycle the calculation sub-steps, change the growth threshold, and re-grow the region according to the changed growth threshold, and fill the grown region to obtain the new area-perimeter ratio of the filled region and the new estimated expected Area-to-perimeter ratio, and the difference between the two, until the difference is the smallest, and the area of the filled area is smaller than the maximum area of the preset growth target, and the area of the filled area is greater than the minimum area of the preset growth target; In the growth sub-step, the growth threshold corresponding to the minimum difference is regarded as the optimal growth threshold, and the growth is performed according to the growth threshold and based on the internal seed point of the vertebral body to obtain the preliminary vertebral body area; The vertebral body positioning step also includes: an optimization sub-step, after obtaining the preliminary vertebral body region, according to the region information of each seed point after region growth, the best vertebral body region is obtained; wherein, the region information includes The position information of the area after the growth of adjacent seed points, and the area difference of the area after the growth of adjacent seed points. 7. A cervical vertebral body positioning device, characterized in that it comprises: The positioning device for the axis of the cervical vertebral body as claimed in claim 4; The seed point selection unit is used to obtain the possible boundary points of the vertebral body according to the gray scale change gradient of the vertebral body axis, filter the possible boundary points according to the third preset condition, and transfer the coordinates of the filtered boundary points along the vertebral body Move in the direction of the axis, and the point corresponding to the new coordinate obtained is the seed point inside the vertebral body; The vertebral body positioning unit is used to obtain the most probable vertebral body region by using the region growing method based on the seed point inside the vertebral body. 8. A cervical intervertebral disc positioning method, characterized in that it comprises: The cervical vertebral body positioning method as claimed in claim 5 or 6; The corner point determination step is to detect the apex of the vertebral body region by using the helical scanning method according to the located cervical vertebral body region; In the step of determining the centerline of the intervertebral disc, two center points are obtained according to the line connecting the vertices of two adjacent vertebral body regions, and the line connecting the two center points is the center line of the intervertebral disc. 9. A cervical intervertebral disc positioning device, characterized in that it comprises: The cervical vertebral body positioning device according to claim 7; The corner point determination unit is used to detect the vertices of the vertebral body region by using the helical scanning method according to the located cervical vertebral body region The intervertebral disc centerline determining unit is configured to obtain two center points according to a line connecting vertices of two adjacent vertebral body regions, and the line connecting the two center points is the intervertebral disc centerline. 10. A magnetic resonance system, characterized by comprising: the cervical vertebral body positioning device as claimed in claim 7 or the cervical intervertebral disc positioning device as claimed in claim 9.

Images