CN114242209B - Medical image preprocessing method and system - Google Patents
Medical image preprocessing method and system Download PDFInfo
- Publication number
- CN114242209B CN114242209B CN202111286692.7A CN202111286692A CN114242209B CN 114242209 B CN114242209 B CN 114242209B CN 202111286692 A CN202111286692 A CN 202111286692A CN 114242209 B CN114242209 B CN 114242209B
- Authority
- CN
- China
- Prior art keywords
- gradient
- medical image
- image
- data set
- processed
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H30/00—ICT specially adapted for the handling or processing of medical images
- G16H30/20—ICT specially adapted for the handling or processing of medical images for handling medical images, e.g. DICOM, HL7 or PACS
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
Landscapes
- Engineering & Computer Science (AREA)
- Medical Informatics (AREA)
- Health & Medical Sciences (AREA)
- Theoretical Computer Science (AREA)
- Software Systems (AREA)
- General Physics & Mathematics (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Physics & Mathematics (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Mathematical Physics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Radiology & Medical Imaging (AREA)
- Epidemiology (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Public Health (AREA)
- Image Analysis (AREA)
- Apparatus For Radiation Diagnosis (AREA)
Abstract
The embodiment of the invention discloses a medical image preprocessing method and a medical image preprocessing system, wherein the method comprises the following steps: constructing a medical image data set, wherein the medical image data set comprises a normal picture data set with normal image quality and an abnormal picture data set with abnormal image quality; performing gradient calculation on the medical image data set to respectively generate a first gradient distribution corresponding to the normal picture data set and a second gradient distribution corresponding to the abnormal picture data set; determining a gradient threshold according to the first gradient distribution and the second gradient distribution; acquiring a medical image to be processed, calculating a third gradient of the medical image to be processed, and acquiring an image category of the medical image to be processed according to the relation between the third gradient and a gradient threshold value; and executing corresponding operation according to the image category. Before the medical picture data are input into the deep learning model for learning, the picture with the whitened or blacked picture quality is screened, noise of the picture with poor quality is filtered, and the training accuracy of the subsequent deep learning model is improved.
Description
Technical Field
The present invention relates to the field of image processing technologies, and in particular, to a medical image preprocessing method and system.
Background
Since DR image quality is affected by equipment parameters, operation levels of operation technicians, physical conditions of patients, poses of patients at photographing, and the like, there are significant differences in brightness, contrast, and the like of picture imaging. When the depth model is trained, the depth learning model is difficult to train due to overlarge quality deviation of part of input pictures, so that the model training accuracy is influenced, and the using effect of the model is influenced.
At present, for the detection of image quality, a deep learning mode is generally adopted to construct a data set of pictures with different quality for classification training. The method has the characteristics of higher training cost and high false positive, namely misjudgment on normal pictures, and can not meet the requirements of high detection rate of low-quality pictures and low misdetection rate of high-quality pictures.
The prior art is therefore still in need of further development.
Disclosure of Invention
Aiming at the technical problems, the embodiment of the invention provides a medical image preprocessing method and a medical image preprocessing system, which can solve the technical problems that in the prior art, pictures are processed through deep learning, misjudgment is easy to exist on normal pictures, and the requirements on high detection rate of low-quality pictures and low misdetection rate of high-quality pictures cannot be met.
A first aspect of an embodiment of the present invention provides a medical image preprocessing method, including:
Constructing a medical image data set, wherein the medical image data set comprises a normal picture data set with normal image quality and an abnormal picture data set with abnormal image quality;
performing gradient calculation on the medical image data set to respectively generate a first gradient distribution corresponding to the normal picture data set and a second gradient distribution corresponding to the abnormal picture data set;
Determining a gradient threshold according to the first gradient distribution and the second gradient distribution;
Acquiring a medical image to be processed, calculating a third gradient of the medical image to be processed, and acquiring an image category of the medical image to be processed according to the relation between the third gradient and a gradient threshold value;
And executing corresponding operation according to the image category.
Optionally, acquiring the medical image to be processed, calculating a third gradient of the medical image to be processed, and acquiring the image class of the medical image to be processed according to the relationship between the third gradient and the gradient threshold value, including:
acquiring a medical image to be processed, calculating a third gradient of the medical image to be processed, and judging whether the third gradient is larger than a gradient threshold value;
If the third gradient is larger than the gradient threshold value, judging the image type of the medical image to be processed as a normal image;
and if the third gradient is smaller than or equal to the gradient threshold value, judging the image type of the medical image to be processed as an abnormal image.
Optionally, when the image class is a normal image, executing a corresponding operation according to the image class, including:
and inputting the medical image to be processed into a deep learning model to obtain a corresponding recognition result.
Optionally, when the image class is an abnormal image, executing a corresponding operation according to the image class, including:
and deleting the medical image to be processed, and not inputting a deep learning model.
Optionally, the determining the gradient threshold according to the first gradient distribution and the second gradient distribution includes:
Respectively counting a first probability distribution map corresponding to the first gradient and a second probability distribution map corresponding to the second gradient by adopting a counting method;
acquiring a first interval in which the maximum probability density of the normal picture data set is located according to the first probability distribution map;
acquiring a second interval in which the maximum probability density of the abnormal picture data set is located according to the second probability distribution diagram;
and determining a gradient threshold according to the first interval and the second interval.
A second aspect of an embodiment of the present invention provides a medical image preprocessing system, the system including: a memory, a processor, and a computer program stored on the memory and executable on the processor, the computer program when executed by the processor implementing the steps of:
Constructing a medical image data set, wherein the medical image data set comprises a normal picture data set with normal image quality and an abnormal picture data set with abnormal image quality;
performing gradient calculation on the medical image data set to respectively generate a first gradient distribution corresponding to the normal picture data set and a second gradient distribution corresponding to the abnormal picture data set;
Determining a gradient threshold according to the first gradient distribution and the second gradient distribution;
Acquiring a medical image to be processed, calculating a third gradient of the medical image to be processed, and acquiring an image category of the medical image to be processed according to the relation between the third gradient and a gradient threshold value;
And executing corresponding operation according to the image category.
Optionally, the computer program when executed by the processor further implements the steps of:
acquiring a medical image to be processed, calculating a third gradient of the medical image to be processed, and judging whether the third gradient is larger than a gradient threshold value;
If the third gradient is larger than the gradient threshold value, judging the image type of the medical image to be processed as a normal image;
and if the third gradient is smaller than or equal to the gradient threshold value, judging the image type of the medical image to be processed as an abnormal image.
Optionally, the computer program when executed by the processor further implements the steps of:
and inputting the medical image to be processed into a deep learning model to obtain a corresponding recognition result.
Optionally, the computer program when executed by the processor further implements the steps of:
Respectively counting a first probability distribution map corresponding to the first gradient and a second probability distribution map corresponding to the second gradient by adopting a counting method;
acquiring a first interval in which the maximum probability density of the normal picture data set is located according to the first probability distribution map;
acquiring a second interval in which the maximum probability density of the abnormal picture data set is located according to the second probability distribution diagram;
and determining a gradient threshold according to the first interval and the second interval.
A third aspect of the embodiment of the present invention provides a non-volatile computer-readable storage medium, where the non-volatile computer-readable storage medium stores computer-executable instructions that, when executed by one or more processors, cause the one or more processors to perform the medical image preprocessing method described above.
In the technical scheme provided by the embodiment of the invention, a medical image data set is constructed, wherein the medical image data set comprises a normal picture data set with normal image quality and an abnormal picture data set with abnormal image quality; performing gradient calculation on the medical image data set to respectively generate a first gradient distribution corresponding to the normal picture data set and a second gradient distribution corresponding to the abnormal picture data set; determining a gradient threshold according to the first gradient distribution and the second gradient distribution; acquiring a medical image to be processed, calculating a third gradient of the medical image to be processed, and acquiring an image category of the medical image to be processed according to the relation between the third gradient and a gradient threshold value; and executing corresponding operation according to the image category. Before the medical picture data are input into the deep learning model for learning, the picture with the whitened or blacked picture quality is screened, noise of the picture with poor quality is filtered, and the training accuracy of the subsequent deep learning model is improved.
Drawings
FIG. 1 is a flow chart of an embodiment of a medical image preprocessing method according to an embodiment of the present invention;
FIG. 2a is a schematic view of an abnormal image of an embodiment of a medical image preprocessing method according to the present invention;
FIG. 2b is a schematic diagram of a normal image of an embodiment of a medical image preprocessing method according to the present invention;
FIG. 3a is a diagram illustrating a histogram distribution of an abnormal image dataset of an embodiment of a medical image preprocessing method according to an embodiment of the present invention when the abnormal image dataset is 3000 images;
FIG. 3b is a schematic diagram showing a histogram distribution of an abnormal image dataset of an embodiment of a medical image preprocessing method according to an embodiment of the present invention when the abnormal image dataset is 5000 images;
FIG. 4a is a diagram illustrating a histogram distribution of a normal image dataset of an embodiment of a medical image preprocessing method according to an embodiment of the present invention when the normal image dataset is 100 images;
FIG. 4b is a diagram illustrating a histogram distribution of 1000 images of a normal image dataset according to an embodiment of a medical image preprocessing method according to the present invention;
FIG. 4c is a diagram illustrating a histogram distribution of a normal image dataset of an embodiment of a medical image preprocessing method according to an embodiment of the present invention when the normal image dataset is 3000 images;
FIG. 4d is a schematic diagram showing a histogram distribution when a normal image dataset of an embodiment of a medical image preprocessing method is 5000 images according to an embodiment of the present invention;
FIG. 5a is a schematic diagram of a to-be-processed image according to an embodiment of a medical image preprocessing method according to the present invention;
Fig. 5b is a schematic diagram of a plum gradient of a picture to be processed according to an embodiment of a medical image preprocessing method according to the present invention;
Fig. 6 is a schematic hardware structure diagram of another embodiment of a medical image preprocessing system according to an embodiment of the present invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to fall within the scope of the invention.
Embodiments of the present invention will be described in detail below with reference to the accompanying drawings.
Referring to fig. 1, fig. 1 is a flowchart illustrating an embodiment of a medical image preprocessing method according to an embodiment of the invention. As shown in fig. 1, includes:
Step S100, constructing a medical image data set, wherein the medical image data set comprises a normal picture data set with normal image quality and an abnormal picture data set with abnormal image quality;
step 200, performing gradient calculation on a medical image data set to respectively generate a first gradient distribution corresponding to a normal picture data set and a second gradient distribution corresponding to an abnormal picture data set;
step S300, determining a gradient threshold according to the first gradient distribution and the second gradient distribution;
Step S400, acquiring a medical image to be processed, calculating a third gradient of the medical image to be processed, and acquiring an image category of the medical image to be processed according to the relation between the third gradient and a gradient threshold value;
and S500, executing corresponding operation according to the image category.
In specific implementation, the embodiment of the invention aims to screen out the picture with the whitened or blacked picture quality before the picture data is input into the deep learning model for learning, filter the noise of the picture with the poor quality and improve the model training accuracy. The blushing or blacking causes unclear texture of lung tissues, which can cause misdiagnosis or missed diagnosis, and the existence of lung diseases can not be judged. The medical image belongs to the scrap standard and should be filtered out during training of the deep learning model. The darkened picture is shown in fig. 2a and the normal picture is shown in fig. 2b.
The medical image of the embodiment of the invention is generally referred to as DR image. And carrying out classification statistics aiming at the quality characteristics of the DR images. Firstly, collecting a data set of a normal picture and a data set of an abnormal picture aiming at the condition that the picture quality is whitened or blacked. And finding out the gradient distribution of the normal picture and the abnormal picture according to the gradient calculation of the picture. And obtaining pictures with high detection rate of low-quality pictures and low false detection rate of high-quality pictures through threshold analysis, and preprocessing the pictures before sending the pictures into a model. The method belongs to a traditional image processing algorithm utilizing image characteristics of waste sheets with poor quality.
Wherein the gradient map is calculated by calculating the difference between the expansion map and the corrosion map of the image. And calculating a gradient map for extracting edge information of the picture. Specific api can call the cv2. Morphyox (img, cv. Morph_gradient, np. Ones ((5, 5), np. Uint 8)), expand similarly to the 'field expansion', expand the highlight region or white part of the image, whose operation result map is larger than the highlight region of the original.
The method provided by the embodiment of the invention has the advantages of high running speed, reliable detection, no need of prior training and the like. Different detection rates and omission rates can be controlled by the threshold.
Further, acquiring the medical image to be processed, calculating a third gradient of the medical image to be processed, and acquiring the image category of the medical image to be processed according to the relation between the third gradient and the gradient threshold value, including:
acquiring a medical image to be processed, calculating a third gradient of the medical image to be processed, and judging whether the third gradient is larger than a gradient threshold value;
If the third gradient is larger than the gradient threshold value, judging the image type of the medical image to be processed as a normal image;
and if the third gradient is smaller than or equal to the gradient threshold value, judging the image type of the medical image to be processed as an abnormal image.
Specifically, assuming that a threshold value a exists, when the gradient of a certain picture is smaller than or equal to a, the picture is judged to be poor in quality, and when the gradient of the certain picture is larger than a, the picture is judged to be good in quality. After the data set is built, determining how many abnormal picture data sets are detected under the standard and how many abnormal picture data sets are not detected under the standard through unifying the standard; also, using the criterion, it is determined how much of the normal picture data set is misjudged to be of low quality, and how much is judged to be normal.
By continuously adjusting the value of a, the behavior of normal and abnormal data sets can be observed. When the detection rate of the abnormal data set is high and the false detection rate of the normal data set is low, it can be determined that a is the optimal threshold at this time.
Further, when the image category is a normal image, executing a corresponding operation according to the image category, including:
and inputting the medical image to be processed into a deep learning model to obtain a corresponding recognition result.
In the implementation, after the image type of the medical image to be processed is determined, if the image is better, a subsequent deep learning model is input.
Further, when the image category is an abnormal image, executing a corresponding operation according to the image category, including:
and deleting the medical image to be processed, and not inputting a deep learning model.
In the implementation, if the image is poor, the image is not input into a subsequent deep learning model for training.
Further, determining a gradient threshold from the first gradient profile and the second gradient profile comprises:
Respectively counting a first probability distribution map corresponding to the first gradient and a second probability distribution map corresponding to the second gradient by adopting a counting method;
acquiring a first interval in which the maximum probability density of the normal picture data set is located according to the first probability distribution map;
acquiring a second interval in which the maximum probability density of the abnormal picture data set is located according to the second probability distribution diagram;
and determining a gradient threshold according to the first interval and the second interval.
In specific implementation, the acquisition method comprises the following steps: assuming that the original graph is 512 x 512, a gradient graph 512 x 512 can be obtained through gradient calculation, the statistical histogram distribution of one gradient graph can be obtained, then the gradient graph of the whole dataset is counted, and the histogram distribution of the whole dataset can be obtained. After the homogenization treatment, a probability distribution map of the gradient map can be obtained.
In the threshold defining method, the maximum probability density of the low quality picture data set is between 5 and 9, and the maximum probability density of the normal picture data set is between 12 and 16. A threshold range is determined. By changing the threshold value one by one, the performance of two data sets, namely, how much data sets with poor quality can be detected, and how much data sets with good quality can be detected by mistake, is observed. When a data set with poor quality is observed, the best threshold is determined when more data sets can be detected, and when a data set with good quality can be detected by mistake.
The invention also provides a specific application embodiment of medical image preprocessing, which comprises the following steps:
constructing normal sample data sets 100, 1000, 3000, 5000 and abnormal sample data sets 100, 1000, 3000, 5000;
Respectively counting gradient distribution histograms of a normal sample data set and an abnormal sample data set under 100, 1000, 3000 and 5000 respectively; wherein the gradient distribution histograms of the abnormal sample data set at 3000 and 5000 are shown in fig. 3a and 3b, respectively, and the gradient distribution histograms of the normal sample data set at 100, 1000, 3000 and 5000 are shown in fig. 4a, 4b, 4c and 4d, respectively.
By analytical comparison, 100,1000,3000,5000 sets of very poor quality, the top 5 gradients were distributed between 5 and 9 (5, 6,7,8, 9);
by analytical comparison, 100,1000,3000,5000 quality-qualified datasets, with the top 5 most gradients distributed between 12-16 (12, 13,14,15, 16);
By analytical comparison, 100,1000,3000,5000 quality-qualified datasets, with the top 5 gradient distributed between 5 and 9 (12, 13,14,15, 16);
The algorithm is designed as follows:
the input picture is sized (512 ), as shown in fig. 5a, generating a gradient map, as shown in fig. 5b;
intercepting a central area of the picture, wherein the width is 400, the length is 400, and the upper left corner point is (50, 50);
counting the value of the first 5 with the highest gradient value, and calculating an average value;
When the threshold value is smaller than a certain value, judging that the picture is waste, and when the threshold value is larger than a certain value, judging that the picture is normal;
carrying out subsequent operation according to the picture types, wherein the waste sheets do not enter training, and the normal sheets enter model training;
the threshold value is adjustable, and the automatic setting can be carried out according to the detection rate and the missed diagnosis rate of the waste pieces.
As can be seen from the above method embodiments, the present invention provides a medical image preprocessing method, by constructing a medical image dataset, the medical image dataset includes a normal picture dataset with normal image quality and an abnormal picture dataset with abnormal image quality; performing gradient calculation on the medical image data set to respectively generate a first gradient distribution corresponding to the normal picture data set and a second gradient distribution corresponding to the abnormal picture data set; determining a gradient threshold according to the first gradient distribution and the second gradient distribution; acquiring a medical image to be processed, calculating a third gradient of the medical image to be processed, and acquiring an image category of the medical image to be processed according to the relation between the third gradient and a gradient threshold value; and executing corresponding operation according to the image category. Before the medical picture data are input into the deep learning model for learning, the picture with the whitened or blacked picture quality is screened, noise of the picture with poor quality is filtered, and the training accuracy of the subsequent deep learning model is improved.
It should be noted that, there is not necessarily a certain sequence between the steps, and those skilled in the art will understand that, in different embodiments, the steps may be performed in different orders, that is, may be performed in parallel, may be performed interchangeably, or the like.
The medical image preprocessing method according to the embodiment of the present invention is described above, and the medical image preprocessing system according to the embodiment of the present invention is described below, referring to fig. 6, fig. 6 is a schematic hardware structure diagram of another embodiment of a medical image preprocessing system according to the embodiment of the present invention, as shown in fig. 6, the system 10 includes: memory 101, processor 102, and a computer program stored on the memory and executable on the processor, which when executed by processor 101, performs the steps of:
Constructing a medical image data set, wherein the medical image data set comprises a normal picture data set with normal image quality and an abnormal picture data set with abnormal image quality;
performing gradient calculation on the medical image data set to respectively generate a first gradient distribution corresponding to the normal picture data set and a second gradient distribution corresponding to the abnormal picture data set;
Determining a gradient threshold according to the first gradient distribution and the second gradient distribution;
Acquiring a medical image to be processed, calculating a third gradient of the medical image to be processed, and acquiring an image category of the medical image to be processed according to the relation between the third gradient and a gradient threshold value;
And executing corresponding operation according to the image category.
Specific implementation steps are the same as those of the method embodiment, and are not repeated here.
Optionally, the computer program when executed by the processor 101 also implements the steps of:
acquiring a medical image to be processed, calculating a third gradient of the medical image to be processed, and judging whether the third gradient is larger than a gradient threshold value;
If the third gradient is larger than the gradient threshold value, judging the image type of the medical image to be processed as a normal image;
and if the third gradient is smaller than or equal to the gradient threshold value, judging the image type of the medical image to be processed as an abnormal image.
Specific implementation steps are the same as those of the method embodiment, and are not repeated here.
Optionally, the computer program when executed by the processor 101 also implements the steps of:
and inputting the medical image to be processed into a deep learning model to obtain a corresponding recognition result.
Specific implementation steps are the same as those of the method embodiment, and are not repeated here.
Optionally, the computer program when executed by the processor 101 also implements the steps of:
and deleting the medical image to be processed, and not inputting a deep learning model.
Specific implementation steps are the same as those of the method embodiment, and are not repeated here.
Optionally, the computer program when executed by the processor 101 also implements the steps of:
Respectively counting a first probability distribution map corresponding to the first gradient and a second probability distribution map corresponding to the second gradient by adopting a counting method;
acquiring a first interval in which the maximum probability density of the normal picture data set is located according to the first probability distribution map;
acquiring a second interval in which the maximum probability density of the abnormal picture data set is located according to the second probability distribution diagram;
and determining a gradient threshold according to the first interval and the second interval.
Specific implementation steps are the same as those of the method embodiment, and are not repeated here.
Embodiments of the present invention provide a non-transitory computer-readable storage medium storing computer-executable instructions for execution by one or more processors, e.g., to perform the method steps S100 through S500 of fig. 1 described above.
By way of example, nonvolatile storage media can include Read Only Memory (ROM), programmable ROM (PROM), electrically Programmable ROM (EPROM), electrically erasable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM), which acts as external cache memory. By way of illustration and not limitation, RAM may be available in many forms such as Synchronous RAM (SRAM), dynamic RAM, (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), SYNCHLINK DRAM (SLDRAM), and Direct Rambus RAM (DRRAM). The disclosed memory components or memories of the operating environment described in embodiments of the present invention are intended to comprise one or more of these and/or any other suitable types of memory.
The above embodiments are only for illustrating the technical solution of the present invention, and are not limiting; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims (3)
1. A medical image preprocessing method, comprising:
Constructing a medical image data set, wherein the medical image data set comprises a normal picture data set with normal image quality and an abnormal picture data set with abnormal image quality;
performing gradient calculation on the medical image data set to respectively generate a first gradient distribution corresponding to the normal picture data set and a second gradient distribution corresponding to the abnormal picture data set;
Determining a gradient threshold according to the first gradient distribution and the second gradient distribution;
Acquiring a medical image to be processed, calculating a third gradient of the medical image to be processed, and acquiring an image category of the medical image to be processed according to the relation between the third gradient and a gradient threshold value;
executing corresponding operation according to the image category;
the obtaining the medical image to be processed, calculating a third gradient of the medical image to be processed, and obtaining the image category of the medical image to be processed according to the relation between the third gradient and the gradient threshold value, including:
acquiring a medical image to be processed, calculating a third gradient of the medical image to be processed, and judging whether the third gradient is larger than a gradient threshold value;
If the third gradient is larger than the gradient threshold value, judging the image type of the medical image to be processed as a normal image;
If the third gradient is smaller than or equal to the gradient threshold value, judging the image type of the medical image to be processed as an abnormal image;
When the image category is a normal image, executing corresponding operation according to the image category, including:
inputting the medical image to be processed into a deep learning model to obtain a corresponding recognition result;
when the image category is an abnormal image, executing corresponding operation according to the image category, including:
deleting the medical image to be processed, and not inputting a deep learning model;
The determining a gradient threshold from the first gradient profile and the second gradient profile comprises:
Respectively counting a first probability distribution map corresponding to the first gradient and a second probability distribution map corresponding to the second gradient by adopting a counting method;
acquiring a first interval in which the maximum probability density of the normal picture data set is located according to the first probability distribution map;
acquiring a second interval in which the maximum probability density of the abnormal picture data set is located according to the second probability distribution diagram;
and determining a gradient threshold according to the first interval and the second interval.
2. A medical image preprocessing system, the system comprising: a memory, a processor, and a computer program stored on the memory and executable on the processor, the computer program when executed by the processor implementing the steps of:
Constructing a medical image data set, wherein the medical image data set comprises a normal picture data set with normal image quality and an abnormal picture data set with abnormal image quality;
performing gradient calculation on the medical image data set to respectively generate a first gradient distribution corresponding to the normal picture data set and a second gradient distribution corresponding to the abnormal picture data set;
Determining a gradient threshold according to the first gradient distribution and the second gradient distribution;
Acquiring a medical image to be processed, calculating a third gradient of the medical image to be processed, and acquiring an image category of the medical image to be processed according to the relation between the third gradient and a gradient threshold value;
executing corresponding operation according to the image category;
the computer program when executed by the processor further performs the steps of:
acquiring a medical image to be processed, calculating a third gradient of the medical image to be processed, and judging whether the third gradient is larger than a gradient threshold value;
If the third gradient is larger than the gradient threshold value, judging the image type of the medical image to be processed as a normal image;
If the third gradient is smaller than or equal to the gradient threshold value, judging the image type of the medical image to be processed as an abnormal image;
the computer program when executed by the processor further performs the steps of:
inputting the medical image to be processed into a deep learning model to obtain a corresponding recognition result;
the computer program when executed by the processor further performs the steps of:
Respectively counting a first probability distribution map corresponding to the first gradient and a second probability distribution map corresponding to the second gradient by adopting a counting method;
acquiring a first interval in which the maximum probability density of the normal picture data set is located according to the first probability distribution map;
acquiring a second interval in which the maximum probability density of the abnormal picture data set is located according to the second probability distribution diagram;
and determining a gradient threshold according to the first interval and the second interval.
3. A non-transitory computer-readable storage medium storing computer-executable instructions that, when executed by one or more processors, cause the one or more processors to perform the medical image preprocessing method of claim 1.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111286692.7A CN114242209B (en) | 2021-11-02 | 2021-11-02 | Medical image preprocessing method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111286692.7A CN114242209B (en) | 2021-11-02 | 2021-11-02 | Medical image preprocessing method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114242209A CN114242209A (en) | 2022-03-25 |
| CN114242209B true CN114242209B (en) | 2024-07-05 |
Family
ID=80743510
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111286692.7A Active CN114242209B (en) | 2021-11-02 | 2021-11-02 | Medical image preprocessing method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114242209B (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110232719A (en) * | 2019-06-21 | 2019-09-13 | 腾讯科技（深圳）有限公司 | Medical image classification method, model training method and server |
| CN111340041A (en) * | 2020-03-13 | 2020-06-26 | 安阳工学院 | License plate recognition method and device based on deep learning |
Family Cites Families (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2009012364A1 (en) * | 2007-07-19 | 2009-01-22 | Nikon Corporation | Device and method for estimating if an image is blurred |
| US9092668B2 (en) * | 2009-07-18 | 2015-07-28 | ABBYY Development | Identifying picture areas based on gradient image analysis |
| US20130044927A1 (en) * | 2011-08-15 | 2013-02-21 | Ian Poole | Image processing method and system |
| CN106228556B (en) * | 2016-07-22 | 2019-12-06 | 北京小米移动软件有限公司 | image quality analysis method and device |
| US10672115B2 (en) * | 2016-12-06 | 2020-06-02 | Siemens Corporation | Weakly supervised anomaly detection and segmentation in images |
| WO2019238804A1 (en) * | 2018-06-13 | 2019-12-19 | Siemens Healthcare Gmbh | Localization and classification of abnormalities in medical images |
| CN110321952B (en) * | 2019-07-02 | 2024-02-09 | 腾讯医疗健康(深圳)有限公司 | Training method of image classification model and related equipment |
| CN110930363A (en) * | 2019-10-29 | 2020-03-27 | 北京临近空间飞行器系统工程研究所 | Method and device for determining sharpness evaluation value of curved-surface blurred image and storage medium |
-
2021
- 2021-11-02 CN CN202111286692.7A patent/CN114242209B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110232719A (en) * | 2019-06-21 | 2019-09-13 | 腾讯科技（深圳）有限公司 | Medical image classification method, model training method and server |
| CN111340041A (en) * | 2020-03-13 | 2020-06-26 | 安阳工学院 | License plate recognition method and device based on deep learning |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114242209A (en) | 2022-03-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11640678B2 (en) | Method for training object detection model and target object detection method | |
| CN108765369B (en) | Method, apparatus, computer device and storage medium for detecting lung nodule | |
| CN111524137B (en) | Cell identification counting method and device based on image identification and computer equipment | |
| CN108108739B (en) | Method and device for detecting image target area, X-ray system and storage medium | |
| CN111369523B (en) | Method, system, equipment and medium for detecting cell stack in microscopic image | |
| CN110956628B (en) | Picture grade classification method, device, computer equipment and storage medium | |
| CN111915541B (en) | Image enhancement processing method, device, equipment and medium based on artificial intelligence | |
| CN117764864B (en) | MRI tumor visual detection method based on image denoising | |
| CN114092450B (en) | Real-time image segmentation method, system and device based on gastroscopy video | |
| CN114241070A (en) | Method and apparatus for removing metal artifacts from CT images and model training | |
| CN110880177A (en) | Image identification method and device | |
| CN114757950A (en) | Ultrasonic image processing method, device and computer readable storage medium | |
| CN116580185A (en) | Defect detection method, system, equipment and medium based on texture | |
| CN119741499B (en) | A method, system and medium for colorectal cancer image segmentation | |
| CN117830387A (en) | Rapid pupil center positioning method for infrared image | |
| CN114242209B (en) | Medical image preprocessing method and system | |
| CN114187303A (en) | Oral cavity image processing method, device, electronic equipment and computer storage medium | |
| CN119339158A (en) | A cholesteatoma image analysis method and system based on deep learning | |
| CN117649415B (en) | Cell balance analysis method based on optical flow diagram detection | |
| CN116342686A (en) | Cell classification counting method and device based on OpenCV image processing | |
| CN113256586B (en) | Method, device, equipment and medium for fuzzy judgment of face image | |
| CN113066108B (en) | Anti-occlusion visual target tracking method and device based on ECO algorithm | |
| CN115661815A (en) | Pathological image classification method and image classification device based on global feature mapping | |
| CN111598184B (en) | DenseNet-based image noise identification method and device | |
| CN117197020A (en) | Mitral valve opening pitch detection method, electronic device, and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |