Computer vision-assisted image enhancement method for reproductive diagnosis

Abstract

The present invention relates to the technical field of image processing, and in particular to a reproductive diagnostic image enhancement method based on computer vision assistance, comprising: obtaining a standard magnetic resonance image of the reproductive system and a to-be-processed magnetic resonance image of the reproductive system, and obtaining a corresponding local tissue image region; calculating a local grayscale signal error and a grayscale equal error probability, and calculating a local grayscale information entropy; obtaining a reproductive system magnetic resonance artifact correction region; obtaining singular values ​​of an image of a reproductive system magnetic resonance artifact correction region, obtaining a correction boundary interval, and calculating a local interval distribution probability of each different interval; obtaining an artifact correction singular value set and reconstructing the image. The present invention effectively improves the quality of the reproductive system magnetic resonance image.

Description

Technical Field

[0001] The present invention relates to the technical field of image processing, and in particular to a reproductive diagnosis image enhancement method based on computer vision assistance. Background Art

[0002] Reproductive diseases are common in young and middle-aged men and women of childbearing age. The main symptoms of patients are lower abdominal pain, etc. In severe cases, it can cause infertility, which has a serious impact on the physical and mental health and normal life of patients. In recent years, with the continuous development and improvement of various assisted reproductive technologies and equipment, and the rapid development of related technologies of medical imaging groups, medical imaging groups can effectively understand the tissue structure characteristics of the patient's lesion area, and further help clinical medical staff understand the patient's condition.

[0003] Magnetic resonance imaging (MRI) technology is a medical image obtained by processing by electronic computers using the phenomenon of nuclear magnetic resonance between atomic nuclei and human tissues in a magnetic field. Unlike other medical imaging methods, MRI has the advantage of less damage to human tissues, and its examination results for reproductive system tissues such as the bladder and uterus are better than other common medical imaging methods such as CT or X-ray. Summary of the invention

[0004] The present invention provides a reproductive diagnosis image enhancement method based on computer vision assistance to solve the existing problems.

[0005] The computer vision-assisted reproductive diagnostic image enhancement method of the present invention adopts the following technical solutions:

[0006] An embodiment of the present invention provides a reproductive diagnostic image enhancement method based on computer vision assistance, the method comprising the following steps:

[0007] Acquire a reproductive system magnetic resonance image of a patient with a reproductive disease, acquire a standard reproductive system magnetic resonance image and a to-be-processed reproductive system magnetic resonance image based on the reproductive system magnetic resonance image, and use the standard reproductive system magnetic resonance image and the to-be-processed reproductive system magnetic resonance image as inputs of a neural network to acquire corresponding local tissue image regions;

[0008] Calculating a local grayscale signal error according to a local tissue image region of a reproductive system magnetic resonance image to be processed and a reproductive system standard magnetic resonance image;

[0009] Divide each local tissue image area according to a preset direction, calculate the grayscale equal error probability in each preset direction according to the local grayscale signal error of each local tissue image, and calculate the local grayscale information entropy of each local tissue image area according to the grayscale equal error probability in each preset direction; obtain a first threshold value according to the local grayscale information entropy of all local tissue image areas, and use the first threshold value to obtain a reproductive system magnetic resonance artifact correction area;

[0010] Each reproductive system MRI artifact correction region image is used as the input of the singular value decomposition algorithm to obtain the singular value of each reproductive system MRI artifact correction region image. The correction boundary interval is obtained according to the singular value of each reproductive system MRI artifact correction region image, and the correction boundary interval is divided into different intervals according to the preset step size. The local interval distribution probability of each different interval is calculated according to the correction boundary interval; the artifact correction singular value set is obtained according to the local interval distribution probability of each different interval; and the reproductive system MRI artifact correction region image is reconstructed according to the artifact correction singular value set.

[0011] Preferably, the step of acquiring a standard MRI image of the reproductive system and a to-be-processed MRI image of the reproductive system according to the MRI image of the reproductive system comprises the following specific steps:

[0012] The reproductive system MRI images with more clinician marks than a preset value are recorded as standard reproductive system MRI images, and the reproductive system MRI images with less clinician marks than a preset value are recorded as reproductive system MRI images to be processed.

[0013] Preferably, the mathematical expression for calculating the local grayscale signal error based on the local tissue image area of ​​the reproductive system magnetic resonance image to be processed and the reproductive system standard magnetic resonance image is:

[0014]

[0015] In the formula, represents the local grayscale signal error of the i-th local area in the reproductive system magnetic resonance image to be processed, represents the mean grayscale signal of the i-th local area in the reproductive system magnetic resonance image to be processed, It represents the mean grayscale signal in the standard MRI image of the reproductive system. It represents the size of the intersection area between the reproductive system MRI image to be processed and the reproductive system standard MRI image. It represents the maximum local area at the corresponding position in the reproductive system magnetic resonance image to be processed and the reproductive system standard magnetic resonance image. is the absolute value function.

[0016] Preferably, the grayscale equal error probability in each preset direction is calculated according to the local grayscale signal error of each local tissue image, and the specific steps include the following:

[0017] Taking each local area in the reproductive system magnetic resonance image to be processed as the central area, the total number of local areas in each preset direction is obtained respectively, which is recorded as the number of first areas of the central area in the corresponding preset direction; the number of local areas whose local grayscale signal errors are the same as that of the central area in each preset direction is recorded as the number of second areas of the central area in the corresponding preset direction; and the ratio of the number of second areas to the number of first areas is recorded as the grayscale equal error probability in each preset direction.

[0018] Preferably, the step of calculating the local grayscale information entropy of each local tissue image region according to the grayscale equal error probability in each preset direction includes the following specific steps:

[0019]

[0020] In the formula, Represents the interval centered on the i-th local area The number of preset directions, It represents the grayscale equal error probability in the jth direction of the reproductive system magnetic resonance image to be processed. represents the logarithmic function with the number 2 as the base, It represents the local grayscale information entropy of the i-th local area in the reproductive system magnetic resonance image to be processed.

[0021] Preferably, the step of obtaining the first threshold value according to the local grayscale information entropy of all local tissue image regions includes the following specific steps:

[0022] The average value of the local grayscale information entropy of all local tissue image regions is calculated and recorded as the first threshold.

[0023] Preferably, the step of obtaining the correction boundary interval according to the singular values ​​of each reproductive system magnetic resonance artifact correction region image comprises the following specific steps:

[0024] The interval formed by the maximum and minimum singular values ​​of the reproductive system magnetic resonance artifact correction region image is used as the correction boundary interval.

[0025] Preferably, the step of calculating the local interval distribution probability of each different interval according to the corrected boundary interval includes the following specific steps:

[0026] The number of singular values ​​falling into each interval is obtained respectively, and the ratio of the number of singular values ​​in each interval to the total number of singular values ​​is recorded as the local interval distribution probability.

[0027] Preferably, the step of calculating the local interval distribution probability of each different interval according to the correction boundary interval; and obtaining the artifact correction singular value set according to the local interval distribution probability of each different interval includes the following specific steps:

[0028] Obtaining the maximum value of all singular values ​​in the reproductive system magnetic resonance artifact correction region image, obtaining the corresponding interval of the maximum value of the local interval distribution probability in the reproductive system magnetic resonance artifact correction image region, and when the maximum singular value exists in the interval corresponding to the maximum local interval distribution probability, putting the singular value into the first set;

[0029] The total number of singular values ​​in the first set is obtained, and the ratio of the total number of singular values ​​in the first set to the total number of all singular values ​​is recorded as a first ratio. When the first ratio is greater than or equal to a preset empirical threshold, the first set is recorded as an artifact-corrected singular value set.

[0030] Preferably, the reconstructing the artifact-corrected regional image of the reproductive system magnetic resonance image according to the artifact-corrected singular value set comprises the following specific steps:

[0031] All other singular value values ​​outside the artifact correction singular value set are set to preset values, and all processed singular values ​​are used as inputs of the image reconstruction algorithm.

[0032] The beneficial effects of the technical solution of the present invention are:

[0033] The present invention uses a neural network to obtain image regions corresponding to different tissue structures in the reproductive system magnetic resonance image, and calculates the corresponding local grayscale signal error through the grayscale signal change characteristics of the image regions of different tissue structures. Compared with directly calculating the entire reproductive system magnetic resonance image, the grayscale signal detail characteristics corresponding to different tissue structure regions of the reproductive system magnetic resonance image can be better obtained. At the same time, considering that the patient will have a certain degree of physiological movement during the magnetic resonance scanning process, these physiological movements present local interference in the reproductive system magnetic resonance image, so the local grayscale information entropy is obtained through the local grayscale signal error, so as to more accurately extract the reproductive system magnetic resonance artifact correction region. Further, since the artifact noise generated by the movement has the characteristics of strong randomness and irregular distribution changes, the present invention analyzes the numerical distribution characteristics of singular values ​​in the reproductive system magnetic resonance artifact correction region, constructs an artifact correction singular value set, screens the effective signal characteristics in the reproductive system magnetic resonance image, and completes the reconstruction of the reproductive system magnetic resonance image. Compared with the image filtering process, it can better obtain effective information in the reproductive system magnetic resonance image and improve the overall quality of the reproductive system magnetic resonance image. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0035] Figure 1 The present invention is a flowchart of the steps of the reproductive diagnosis image enhancement method based on computer vision assistance. DETAILED DESCRIPTION

[0036] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of the specific implementation, structure, features and effects of the reproductive diagnostic image enhancement method based on computer vision assistance proposed by the present invention in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0037] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0038] The specific scheme of the reproductive diagnosis image enhancement method based on computer vision assistance provided by the present invention is described in detail below with reference to the accompanying drawings.

[0039] See also Figure 1 , which shows a flowchart of a method for enhancing reproductive diagnostic images based on computer vision assistance provided by an embodiment of the present invention, the method comprising the following steps:

[0040] Step S001: Acquire a reproductive system magnetic resonance image of a patient with reproductive disease, acquire a standard reproductive system magnetic resonance image and a to-be-processed reproductive system magnetic resonance image based on the reproductive system magnetic resonance image, and use the standard reproductive system magnetic resonance image and the to-be-processed reproductive system magnetic resonance image as inputs of a neural network to acquire corresponding local tissue image regions.

[0041] Medical imaging equipment is used to perform MRI scans on patients with reproductive diseases using T2-weighted imaging, and the scan results are recorded as reproductive system magnetic resonance images.

[0042] Furthermore, when more than three clinicians believe that the clarity of the reproductive system MRI image does not affect the judgment of clinical diseases, the image is recorded as a standard reproductive system MRI image, otherwise it is recorded as a reproductive system MRI image to be processed.

[0043] It should be noted that due to the different internal structures of the human reproductive system, the resonance intensity of electromagnetic signals in different tissue structures under the magnetic field environment will also be different, and the details of different tissue regions in the obtained MRI images will also be different. Therefore, it is necessary to first process the reproductive system magnetic resonance images to obtain the local tissue image areas corresponding to different tissue structures.

[0044] Specifically, the reproductive system magnetic resonance image to be processed and the standard reproductive system magnetic resonance image are used as input of the U-Net neural network model to obtain image areas with different signal intensities in the reproductive system magnetic resonance image and the standard reproductive system magnetic resonance image, and the segmented images are arranged in order from large to small according to the grayscale signal mean of the pixels in the area to obtain different local areas.

[0045] The U-Net network is trained using the reproductive system magnetic resonance images. The U-Net network model structure and specific training techniques are well-known technologies and will not be described in detail here. It should be noted that the U-Net network model used is only a technical means for obtaining the segmentation of different tissue regions in the reproductive system magnetic resonance images. No specific restrictions are made here, and the implementer can choose according to the actual situation.

[0046] Step S002: Calculating a local grayscale signal error according to the local tissue image region of the reproductive system magnetic resonance image to be processed and the reproductive system standard magnetic resonance image.

[0047] It should be noted that the local structures of the various tissues and organs in the reproductive system are different, and the MRI performance of different tissues will also be different. Due to the uneven magnetic field strength during the MRI scanning process, the image will be degenerated during the scanning process, and the signal intensity in the MRI image will be different from the standard intensity, resulting in unclear MRI images of the reproductive system. Therefore, it is necessary to combine the standard MRI images of the reproductive system for processing.

[0048] Specifically, different local areas in the reproductive system magnetic resonance image to be processed are compared and calculated with the standard nuclear magnetic resonance image of the reproductive system to obtain the local grayscale signal error. The corresponding calculation formula is:

[0049]

[0050] In the formula, represents the local grayscale signal error of the i-th local area in the reproductive system magnetic resonance image to be processed, represents the mean grayscale signal of the i-th local area in the reproductive system magnetic resonance image to be processed, It represents the mean grayscale signal in the standard MRI image of the reproductive system. It represents the size of the intersection area between the reproductive system MRI image to be processed and the reproductive system standard MRI image. It represents the maximum local area at the corresponding position in the reproductive system magnetic resonance image to be processed and the reproductive system standard magnetic resonance image. is the absolute value function.

[0051] The above formula can be used to calculate the local grayscale signal error of the i-th local area in the reproductive system magnetic resonance image to be processed. The above calculation is performed on different local areas in the reproductive system magnetic resonance image to be processed to obtain the local grayscale signal errors of all different local areas. Due to certain differences in the details of the reproductive system structure composition between different human bodies, the larger the communication area between the local area of ​​the reproductive system magnetic resonance image to be processed and the standard nuclear magnetic resonance image of the reproductive system, the larger the image proportion in the largest local area will be. At the same time, if the local grayscale signal difference between the two different images is smaller, the calculated local grayscale signal error is smaller, indicating that the error of the reproductive system magnetic resonance image to be processed is also relatively small at this time. On the contrary, if the local grayscale signal error is larger, it means that there is a large error between the local area and the standard normal magnetic resonance image.

[0052] Step S003: Divide each local tissue image area according to a preset direction, calculate the grayscale equal error probability in each preset direction according to the local grayscale signal error of each local tissue image, and calculate the local grayscale information entropy of each local tissue image area according to the grayscale equal error probability in each preset direction; obtain a first threshold value according to the local grayscale information entropy of all local tissue image areas, and use the first threshold value to obtain the reproductive system magnetic resonance artifact correction area.

[0053] It should be noted that the MRI magnetic resonance imaging process has the disadvantages of long scanning time and slow imaging compared to other imaging methods. At the same time, patients who are in a fixed scanning position for a long time will have involuntary movements such as swallowing and coughing during the scanning process due to psychological fear and physical discomfort. These physiological and physical involuntary movements will cause artifacts of varying degrees in the MRI images, resulting in errors in the final obtained reproductive system magnetic resonance images to be processed. Therefore, it is necessary to use local grayscale signals to misprocess the reproductive system magnetic resonance images to be processed.

[0054] Specifically, during the magnetic resonance imaging process, since different tissue regions have different magnetic resonance nuclear resonance signal intensities, the error region in the reproductive system magnetic resonance image to be processed is firstly acquired.

[0055]

[0056] In the formula, Represents the interval centered on the i-th local area 8 different directions, It represents the grayscale equal error probability in the jth direction of the reproductive system magnetic resonance image to be processed. represents the logarithmic function with the number 2 as the base, It represents the local grayscale information entropy of the i-th local area in the reproductive system magnetic resonance image to be processed.

[0057] The above formula can be used to calculate the local grayscale information entropy of different local areas in the reproductive system magnetic resonance image to be processed. If the patient has involuntary physiological movements during the MRI scanning process, artifacts will appear in the reproductive system magnetic resonance image to be processed. Therefore, the smaller the grayscale error probability calculated in the 8 different directions of the center of the i-th local area, the more chaotic the local information of the reproductive system magnetic resonance image, and the larger the calculated local grayscale information entropy, the higher the possibility of artifacts in the local area of ​​the reproductive system magnetic resonance image.

[0058] The calculation method of the grayscale equal error probability in the j-th direction in the reproductive system magnetic resonance image to be processed is:

[0059] The total number of local regions in the jth direction with the i local region as the center in the reproductive system magnetic resonance image to be processed is obtained, which is recorded as the first region number. At the same time, the number of local regions in the jth direction in the reproductive system magnetic resonance image to be processed with the same local grayscale signal error as that of the i local region is obtained, which is recorded as the second region number. The ratio of the second region number to the first region number is recorded as the grayscale equal error probability in the jth direction in the reproductive system magnetic resonance image to be processed.

[0060] The local grayscale information entropy of all different local areas in the reproductive system magnetic resonance image to be processed is obtained, and the average value thereof is recorded as a first threshold, and all local areas with a value greater than the first threshold are recorded as reproductive system magnetic resonance artifact correction areas.

[0061] Step S004: using each reproductive system MRI artifact correction region image as the input of the singular value decomposition algorithm, obtaining the singular value of each reproductive system MRI artifact correction region image, obtaining the correction boundary interval according to the singular value of each reproductive system MRI artifact correction region image, dividing the correction boundary interval into different intervals according to a preset step size, calculating the local interval distribution probability of each different interval according to the correction boundary interval; obtaining an artifact correction singular value set according to the local interval distribution probability of each different interval; and reconstructing the reproductive system MRI artifact correction region image according to the artifact correction singular value set.

[0062] It should be noted that during the MRI scanning process, patients in a normal breathing state will have motion artifacts due to physiological movements such as breathing and swallowing. Normally, normal physiological activities occur in local tissue areas. For example, patients who have been in a fixed position for a long time have muscle spasms, and the images of the spasm area will cause local artifacts in the MRI image. At the same time, considering that the normal physiological movement of patients only involves part of the body's movement tissue, it is necessary to dynamically process the local artifact area.

[0063] Specifically, the image of the reproductive system magnetic resonance artifact correction region is used as the input of the singular value decomposition (SVD) algorithm to obtain the singular values ​​of the reproductive system magnetic resonance artifact correction region image. It should be noted that singular value decomposition is a commonly used algorithm for obtaining singular values ​​of an image, and its specific calculation process is a well-known technology and will not be repeated here.

[0064] When decomposing the reproductive system magnetic resonance artifact correction area images, due to the long MRI scanning time, there will be motion artifact noise in the acquired reproductive system magnetic resonance images. Since these motion artifact noises represent different meanings, the singular values ​​corresponding to these areas will have numerical differences from the singular values ​​corresponding to the normal areas.

[0065] It should be noted that the singular values ​​and corresponding vectors of the reproductive system MRI artifact correction region image obtained after SVD decomposition represent different change characteristics of the reproductive system MRI artifact correction region image stretching, scaling, etc. Under normal circumstances, the changes formed by the superposition of motion artifact noise are obviously unreasonable, and the numerical distribution of most motion artifact noise singular values ​​is relatively scattered.

[0066] When motion noise artifacts appear in the reproductive system MRI artifact correction region image, the singular value calculated in the motion noise artifact region is relatively small because the motion noise artifact is usually scattered and meaningless. The normal image region reflects the changes in the tissue structure of the reproductive system, and the singular value calculated in the corresponding image region is relatively large. Therefore, the motion noise artifacts superimposed in the reproductive system MRI artifact correction region image are separated from the singular value values ​​of the reproductive system MRI artifact correction region image.

[0067] Specifically, the maximum and minimum values ​​of the singular values ​​of the reproductive system magnetic resonance artifact correction area image are taken as the correction boundary interval, and the correction boundary interval is divided according to a step size of S to obtain different intervals, where S takes an empirical value of 2. In actual situations, implementers can set it according to specific circumstances.

[0068] The number of singular values ​​falling into each interval is obtained respectively, and the ratio of the number of singular values ​​in each interval to the total number of singular values ​​is recorded as the local interval distribution probability. The artifact correction singular value set is initialized, and among all the singular values ​​in the artifact correction area image of the reproductive system magnetic resonance, the singular value with the largest singular value and the largest local interval distribution probability corresponding to the singular value is first selected, and the selected singular value is merged into the artifact correction singular value set.

[0069] Repeat the selection step until the ratio of the number of singular values ​​in the artifact correction singular value set to the total number of singular values ​​is greater than or equal to the preset empirical threshold, wherein the preset empirical threshold is 0.9, and in actual situations, the implementer can set it according to the specific situation. The artifact correction singular value set can be used to screen and obtain the singular values ​​of the normal image in the reproductive system magnetic resonance image. Therefore, all other singular values ​​outside the artifact correction singular value set are set to 0, and the singular values ​​after the numerical value is reset are used as the input of the image reconstruction algorithm to obtain the reconstructed reproductive system magnetic resonance artifact correction area image. Among them, the specific process of reconstructing the image using singular values ​​is a well-known technology and will not be repeated here.

[0070] By statistically analyzing the numerical distribution of singular values ​​in the reproductive system MRI artifact correction area images, the artifact manifestation characteristics of the reproductive system MRI artifact correction area images are effectively obtained, and the images are reconstructed using the reset singular value pairs, thereby effectively eliminating the motion artifact change noise of the reproductive system MRI artifact correction area images.

[0071] So far, the present invention is completed.

[0072] To summarize, in an embodiment of the present invention, a reproductive system magnetic resonance image is acquired, and the detail change characteristics of different regions in the reproductive system magnetic resonance image are analyzed to acquire the image region that needs to be corrected. By acquiring the change characteristics of motion artifact noise in the reproductive system magnetic resonance image region, the motion artifact noise is extracted to filter out the influence of the motion artifact noise in the reproductive system magnetic resonance image, thereby providing accurate and reliable image support for the clinical identification of reproductive system diseases.

[0073] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A computer vision-assisted reproductive diagnostic image enhancement method, characterized in that: The method comprises the following steps: Acquire a reproductive system magnetic resonance image of a patient with a reproductive disease, acquire a standard reproductive system magnetic resonance image and a to-be-processed reproductive system magnetic resonance image based on the reproductive system magnetic resonance image, and use the standard reproductive system magnetic resonance image and the to-be-processed reproductive system magnetic resonance image as inputs of a neural network to acquire corresponding local tissue image regions; Calculating a local grayscale signal error according to a local tissue image region of a reproductive system magnetic resonance image to be processed and a reproductive system standard magnetic resonance image; Divide each local tissue image area according to a preset direction, calculate the grayscale equal error probability in each preset direction according to the local grayscale signal error of each local tissue image, and calculate the local grayscale information entropy of each local tissue image area according to the grayscale equal error probability in each preset direction; obtain a first threshold value according to the local grayscale information entropy of all local tissue image areas, and use the first threshold value to obtain a reproductive system magnetic resonance artifact correction area; Using each reproductive system magnetic resonance artifact correction region image as the input of the singular value decomposition algorithm, obtaining the singular value of each reproductive system magnetic resonance artifact correction region image, obtaining the correction boundary interval according to the singular value of each reproductive system magnetic resonance artifact correction region image, dividing the correction boundary interval into different intervals according to a preset step size, and calculating the local interval distribution probability of each different interval according to the correction boundary interval; obtaining an artifact correction singular value set according to the local interval distribution probability of each different interval; and reconstructing the reproductive system magnetic resonance artifact correction region image according to the artifact correction singular value set; The mathematical expression for calculating the local grayscale signal error based on the local tissue image area of ​​the reproductive system magnetic resonance image to be processed and the reproductive system standard magnetic resonance image is: In the formula, represents the local grayscale signal error of the i-th local area in the reproductive system magnetic resonance image to be processed, represents the mean grayscale signal of the i-th local area in the reproductive system magnetic resonance image to be processed, It represents the mean grayscale signal in the standard MRI image of the reproductive system. It represents the size of the intersection area between the reproductive system MRI image to be processed and the reproductive system standard MRI image. It represents the maximum local area at the corresponding position in the reproductive system magnetic resonance image to be processed and the reproductive system standard magnetic resonance image. is the absolute value function; The specific steps of calculating the grayscale equal error probability in each preset direction according to the local grayscale signal error of each local tissue image include the following: Taking each local area in the reproductive system magnetic resonance image to be processed as the central area, respectively obtaining the total number of local areas in each preset direction, recording it as the number of first areas of the central area in the corresponding preset direction, recording the number of local areas with the same local grayscale signal error as the central area in each preset direction as the number of second areas of the central area in the corresponding preset direction, and recording the ratio of the number of second areas to the number of first areas as the grayscale equal error probability in each preset direction; The specific steps of calculating the local grayscale information entropy of each local tissue image area according to the grayscale equal error probability in each preset direction include the following: In the formula, Represents the interval centered on the i-th local area The number of preset directions, It represents the grayscale equal error probability in the jth direction of the reproductive system magnetic resonance image to be processed. represents the logarithmic function with the number 2 as the base, It represents the local grayscale information entropy of the i-th local area in the reproductive system magnetic resonance image to be processed.

2. The computer vision-assisted reproductive diagnostic image enhancement method according to claim 1, characterized in that: The specific steps of obtaining the standard MRI image of the reproductive system and the MRI image of the reproductive system to be processed according to the MRI image of the reproductive system are as follows: The reproductive system MRI images with more clinician marks than a preset value are recorded as standard reproductive system MRI images, and the reproductive system MRI images with less clinician marks than a preset value are recorded as reproductive system MRI images to be processed.

3. The computer vision-assisted reproductive diagnostic image enhancement method according to claim 1, characterized in that: The specific steps of obtaining the first threshold value according to the local grayscale information entropy of all local tissue image regions are as follows: The average value of the local grayscale information entropy of all local tissue image regions is calculated and recorded as the first threshold.

4. The computer vision-assisted reproductive diagnostic image enhancement method according to claim 1, characterized in that: The specific steps of obtaining the correction boundary interval according to the singular values ​​of each reproductive system magnetic resonance artifact correction region image include the following: The interval formed by the maximum and minimum singular values ​​of the reproductive system magnetic resonance artifact correction region image is used as the correction boundary interval.

5. The computer vision-assisted reproductive diagnostic image enhancement method according to claim 4, characterized in that: The specific steps of calculating the local interval distribution probability of each different interval according to the corrected boundary interval are as follows: The number of singular values ​​falling into each interval is obtained respectively, and the ratio of the number of singular values ​​in each interval to the total number of singular values ​​is recorded as the local interval distribution probability.

6. The computer vision-assisted reproductive diagnostic image enhancement method according to claim 1, characterized in that: The method of calculating the local interval distribution probability of each different interval according to the correction boundary interval; and obtaining the artifact correction singular value set according to the local interval distribution probability of each different interval includes the following specific steps: Obtaining the maximum value of all singular values ​​in the reproductive system magnetic resonance artifact correction region image, obtaining the corresponding interval of the maximum value of the local interval distribution probability in the reproductive system magnetic resonance artifact correction image region, and when the maximum singular value exists in the interval corresponding to the maximum local interval distribution probability, putting the singular value into the first set; The total number of singular values ​​in the first set is obtained, and the ratio of the total number of singular values ​​in the first set to the total number of all singular values ​​is recorded as a first ratio. When the first ratio is greater than or equal to a preset empirical threshold, the first set is recorded as an artifact-corrected singular value set.

7. The computer vision-assisted reproductive diagnostic image enhancement method according to claim 6, characterized in that: The specific steps of reconstructing the artifact-corrected regional image of the reproductive system magnetic resonance according to the artifact-corrected singular value set are as follows: All other singular value values ​​outside the artifact correction singular value set are set to preset values, and all processed singular values ​​are used as inputs of the image reconstruction algorithm.

Images