Auxiliary analysis method and system for infertility ultrasound examination images
By calculating the gradient feature vector and the final feature vector for the pelvic ultrasound image, combined with the central position of the uterus, using the SVW support vector machine to distinguish and enhance the fallopian tube disease area, the problem of inability to effectively enhance the characteristics of fallopian tube disease in the prior art is solved. Improves the diagnostic accuracy of infertility ultrasound examination.
Patent Information
- Application Number
- CN202411489120.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-24
- Publication Date
- 2024-12-17
- Estimated Expiration
- 2044-10-24
AI Technical Summary
The existing image enhancement algorithm cannot effectively enhance the characteristics of the fallopian tubes, resulting in a decrease in the accuracy of the diagnosis results of infertility ultrasound examination.
By obtaining pelvic ultrasound images, it is divided into several cell units. The gradient feature vector and final feature vector of each cell unit are calculated. Combined with the central position of the uterus, the SVW support vector machine is used to distinguish and enhance the fallopian tube disease area.
It improves the enhancement effect on fallopian tube disease areas, enhances the diagnostic accuracy of infertility ultrasound examination, and reduces the possibility of wrong diagnosis.
Smart Images
Figure CN119006468B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of image processing, and in particular to an auxiliary analysis method and system for infertility ultrasonic examination images. Background Art
[0002] Infertility is a relatively common disease. The main cause of infertility is tubal infertility. When diagnosing infertility, ultrasound examination is often used to collect ultrasound images of the pelvis, and then the doctor diagnoses the patient's condition through analysis of the ultrasound images. Doctors mainly observe pelvic ultrasound images with the naked eye, and judge the patient's condition by observing whether there are symptoms of tubal hypertrophy caused by tubal obstruction and water accumulation in the pelvic ultrasound images. In this process, the patient's diagnosis result is greatly affected by the quality of the pelvic ultrasound image. If the imaging quality of the pelvic ultrasound image at the fallopian tube is not high, it may lead to an erroneous diagnosis. Therefore, it is necessary to enhance the diseased area of the fallopian tube position in the pelvic ultrasound image. When enhancing the diseased area at the fallopian tube position, the existing image enhancement algorithm cannot enhance the diseased characteristics of the fallopian tube, resulting in an insignificant enhancement effect, and then the possible diseased area cannot be highlighted during the doctor's diagnosis process, reducing the accuracy of the diagnosis results of infertility ultrasound examination. Summary of the invention
[0003] The present invention provides an auxiliary analysis method and system for infertility ultrasound examination images to solve the problem that the existing image enhancement algorithm cannot enhance the characteristics of fallopian tube symptoms. The technical solution adopted is as follows:
[0004] The present invention proposes an auxiliary analysis method for infertility ultrasound examination images, which comprises the following steps:
[0005] Acquire a pelvic ultrasound image; divide the pelvic ultrasound image into a number of cell units; obtain a gradient feature vector of each cell unit according to the gradient of the pixel points in each cell unit; the gradient feature vector includes a number of gradient feature values;
[0006] According to the difference of the gradient eigenvalues in the gradient eigenvalue vector of each cell unit, a preliminary adjustment value of each gradient eigenvalue in the gradient eigenvalue vector of each cell unit is obtained; the preliminary adjustment values of all the gradient eigenvalues in the gradient eigenvalue vector of each cell unit are sorted to obtain the gradient adjustment vector of each cell unit;
[0007] According to the difference between the gradient adjustment vectors of each cell unit and the surrounding cell units, the gradient correction vector of each cell unit is obtained;
[0008] The center position of the uterus is obtained according to the pelvic ultrasound image; the final feature vector of each cell unit is obtained according to the center position of the uterus and the gradient correction vector of each cell unit;
[0009] According to the final feature vectors of all cell units in the pelvic ultrasound image, the examination results of the infertility ultrasound image are obtained.
[0010] Furthermore, the step of obtaining the gradient feature vector of each cell unit according to the gradient of the pixel points in each cell unit includes the following specific steps:
[0011] Divide 360 degrees into equal parts direction intervals, and each direction interval is recorded as a bin direction; among them, is the preset number of bin directions; get the value of each cell in A directional gradient histogram is obtained in each bin direction, wherein the directional gradient histogram contains the gradient eigenvalues in each bin direction, and all the gradient eigenvalues are sorted in the order of the bin direction corresponding to each gradient eigenvalue to obtain the gradient eigenvector of each cell unit.
[0012] Furthermore, the step of obtaining a preliminary adjustment value of each gradient eigenvalue in the gradient eigenvalue vector of each cell unit according to the difference of the gradient eigenvalue in the gradient eigenvalue vector of each cell unit includes the following specific steps:
[0013] According to the difference of the gradient eigenvalues in the gradient eigenvector of each cell unit, the disease degree of the gradient eigenvector of each cell unit is obtained;
[0014] According to the gradient eigenvalue in the gradient eigenvector of each cell unit, the necessity of maintaining each gradient eigenvalue in the gradient eigenvector of each cell unit is obtained;
[0015] According to the disease degree of the gradient feature vector of each cell unit and each gradient feature value in the gradient feature vector of each cell unit and the necessity of maintaining the gradient feature value, a preliminary adjustment value of each gradient feature value in the gradient feature vector of each cell unit is obtained.
[0016] Furthermore, according to the disease degree of the gradient feature vector of each cell unit and each gradient feature value in the gradient feature vector of each cell unit and the necessity of maintaining the gradient feature value, a preliminary adjustment value of each gradient feature value in the gradient feature vector of each cell unit is obtained, and the specific calculation method included is as follows:
[0017] ;
[0018] In the formula, For the The gradient feature vector of the cell unit The initial adjustment value of the gradient eigenvalue; For the The sum of all gradient eigenvalues in the gradient eigenvector of a cell; For the The maximum gradient eigenvalue in the gradient eigenvector of the cell unit; is the number of bin directions; For the The first cell in the gradient feature vector Gradient eigenvalues; is the normalization function; For the The severity of the gradient feature vector of each cell; For the The gradient feature vector of the cell unit The necessity of maintaining the gradient eigenvalues.
[0019] Furthermore, the step of sorting the preliminary adjustment values of all gradient feature values in the gradient feature vector of each cell unit to obtain the gradient adjustment vector of each cell unit includes the following specific steps:
[0020] The preliminary adjustment values of all gradient eigenvalues in the gradient eigenvector of each cell unit are sorted according to the order of the gradient eigenvalues in the gradient eigenvector to obtain the gradient adjustment vector of each cell unit.
[0021] Furthermore, the step of obtaining the gradient correction vector of each cell unit according to the difference between the gradient adjustment vectors of each cell unit and the surrounding cell units includes the following specific steps:
[0022] Record any cell as the target cell, and record all cells within the eight neighborhoods of the target cell as reference cells of the target cell; calculate the The cell unit and the The modulus of the difference vector of the gradient adjustment vector of each reference cell unit of the cell unit, the normalized result of the smallest modulus and the The product of the gradient adjustment vectors of the cell units is recorded as The gradient correction vector of the cell unit.
[0023] Furthermore, the step of obtaining the center position of the uterus according to the pelvic ultrasound image includes the following specific steps:
[0024] The pelvic ultrasound image is divided into two types of areas by the Otsu threshold method, and the area with the largest mean grayscale value of the pixels in the area is recorded as the main area; a two-dimensional rectangular coordinate system is established with the lower left corner of the pelvic ultrasound image as the origin, the horizontal direction as the x-axis, and the vertical direction as the y-axis; the mean of the x-coordinates corresponding to all the pixels in the main area is recorded as the horizontal coordinate of the uterus center; the mode of the y-coordinates corresponding to all the pixels in the main area is recorded as the vertical coordinate of the uterus center; the center position of the uterus is obtained according to the horizontal coordinate and the vertical coordinate of the uterus center.
[0025] Furthermore, the final feature vector of each cell unit is obtained according to the center position of the uterus and the gradient correction vector of each cell unit, and the specific steps include the following:
[0026] For any cell unit, the distance from the central pixel of the cell unit to the center of the uterus is recorded as the characteristic distance of the cell unit;
[0027] The first The normalized result of the feature distance of the cell unit is the same as that of the The product of the gradient correction vectors of the cell units is recorded as The final feature vector of the cell.
[0028] Furthermore, the method of obtaining the examination result of the infertility ultrasound image according to the final feature vectors of all cell units in the pelvic ultrasound image includes the following specific steps:
[0029] The final feature vectors of all cell units in the pelvic ultrasound image are input into the SVW support vector machine to obtain the ultrasound imaging examination results of infertility; the ultrasound imaging examination results of infertility include normal fallopian tube images or abnormal fallopian tube images.
[0030] The present invention also proposes an auxiliary analysis system for infertility ultrasound examination images, including a memory, a processor, and a computer program stored in the memory and running on the processor, and the processor executes the computer program stored in the memory to implement the steps of the aforementioned auxiliary analysis method for infertility ultrasound examination images.
[0031] The beneficial effects of the technical solution of the present invention are as follows: the present invention obtains the gradient eigenvalue of each cell unit in each bin direction, and according to the difference of the gradient eigenvalues in each bin direction in a single cell unit, judges the possibility that each cell unit is located in the fallopian tube disease area, and preliminarily distinguishes the normal area from the disease area; the reliability of distinguishing the normal area from the disease area only by the difference between the gradient eigenvalues in a single cell unit is not high, so the similarity between the gradient eigenvalues in multiple cell units is combined to further distinguish the normal area from the disease area, and improves the accuracy of the gradient eigenvector enhancement; since the pelvic ultrasound image contains multiple tissue areas, in order to avoid the interference of other tissue areas, the distance relationship between the cell unit and the center position of the uterus is constructed to exclude the influence of other tissue areas on the enhancement effect. The final eigenvectors of all cell units are input into the SVW support vector machine to obtain the ultrasound imaging results of infertility, which improves the enhancement effect of the possible disease area in the pelvic ultrasound image, thereby assisting doctors in performing ultrasound examinations and indirectly improving the accuracy of doctors' examinations. BRIEF DESCRIPTION OF THE DRAWINGS
[0032] 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.
[0033] Figure 1 The present invention is a flowchart of the steps of an auxiliary analysis method for infertility ultrasound examination images. DETAILED DESCRIPTION
[0034] 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 auxiliary analysis method and system of infertility ultrasound examination images proposed by the present invention, its specific implementation method, structure, characteristics and effects, 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.
[0035] 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.
[0036] The specific scheme of the auxiliary analysis method and system of infertility ultrasound examination images provided by the present invention is described in detail below with reference to the accompanying drawings.
[0037] See also Figure 1 , which shows a flowchart of a method for auxiliary analysis of infertility ultrasound examination images provided by an embodiment of the present invention, the method comprising the following steps:
[0038] Step S001: Acquire a pelvic ultrasound image; divide the pelvic ultrasound image into a number of cell units; and obtain a gradient feature vector of each cell unit according to the gradient of the pixel points in each cell unit.
[0039] The purpose of this embodiment is to enhance the ultrasonic image of the pelvic region according to the disease characteristics of the fallopian tube, so it is necessary to collect the pelvic ultrasonic image.
[0040] Specifically, the original pelvic ultrasound image of the patient is obtained in the ultrasound department of the hospital; the original pelvic ultrasound image is processed by a gamma correction method to obtain a pelvic ultrasound image; the gamma correction method is a well-known technology, and the specific method will not be introduced here.
[0041] It should be noted that this embodiment mainly adjusts the HOG algorithm to adjust the pixel gradient to achieve the purpose of enhancing the pelvic ultrasound image, so it is necessary to segment the pelvic ultrasound image into several cell units and obtain the directional gradient histogram of each cell unit.
[0042] Specifically, using The Sobel operator is used to obtain the gradient of each pixel in the pelvic ultrasound image, and the pelvic ultrasound image is divided into several pixels with a side length of Each block is recorded as a cell unit, and 360 degrees are divided into direction intervals, and each direction interval is recorded as a bin direction; among them, is the preset single block side length, , is the preset number of bin directions, This embodiment is described by taking this as an example; obtaining each cell unit in The directional gradient histogram in each bin direction is obtained, and the directional gradient histogram of each cell unit contains the gradient eigenvalue in each bin direction. All gradient eigenvalues are sorted in the order of the bin direction corresponding to each gradient eigenvalue to obtain the gradient eigenvector of each cell unit; the Sobel operator obtains the gradient of the pixel point, the overlap segmentation method and the directional gradient histogram acquisition method are well-known technologies, and the specific method will not be introduced here.
[0043] Step S002: obtaining a preliminary adjustment value of each gradient eigenvalue in the gradient eigenvalue vector of each cell unit according to the difference of the gradient eigenvalue in the gradient eigenvalue vector of each cell unit; and sorting the preliminary adjustment values of all the gradient eigenvalues in the gradient eigenvalue vector of each cell unit to obtain the gradient adjustment vector of each cell unit.
[0044] It should be noted that this embodiment mainly enhances the diseased area of the fallopian tube in the pelvic ultrasound image and suppresses the gradient of the normal area, thereby achieving the purpose of enhancing the pelvic ultrasound image. The disease characteristics of the fallopian tube are mainly manifested in the end hypertrophy caused by the blockage of the fallopian tube and the accumulation of water. In the pelvic ultrasound image, the diseased area of the fallopian tube is compared with the normal area. The edge of the normal area of the fallopian tube is relatively straight, while the edge of the diseased area of the fallopian tube is locally curved and uneven. Therefore, for the cell unit belonging to the normal area of the fallopian tube, since the edge of the normal area is flat and the gradient direction is relatively similar, the data in the directional gradient histogram of the cell unit in the normal area is mainly concentrated in one bin direction, and the data in other bin directions is relatively low; while for the cell unit belonging to the diseased area of the fallopian tube, since the edge of the diseased area is locally curved and uneven, the gradient direction is different, so the data in the directional gradient histogram of the cell unit in the diseased area is scattered in multiple bin directions, so the normal area and the diseased area of the fallopian tube are preliminarily distinguished according to this feature.
[0045] Specifically, The gradient feature vector of the cell unit The calculation method for the initial adjustment value of the gradient eigenvalue is:
[0046] ;
[0047] In the formula, For the The gradient feature vector of the cell unit The initial adjustment value of the gradient eigenvalue; For the The sum of all gradient eigenvalues in the gradient eigenvector of a cell; For the The maximum gradient eigenvalue in the gradient eigenvector of the cell unit; is the number of bin directions; For the The first cell in the gradient feature vector Gradient eigenvalues; is a linear normalization function, and the normalized object is all the gradient eigenvalues in the gradient eigenvector of all cell units. .
[0048] It should be noted that Indicates The smaller the value of the gradient feature vector of the cell unit, the greater the severity of the disease. The more likely a cell is to belong to the normal area, The smaller the adjusted gradient value of the pixel points in the corresponding area of the cell unit in the pelvic ultrasound image is; Indicates The gradient feature vector of the cell unit The necessity of maintaining the gradient eigenvalue is that the smaller the value is, the The first cell in the gradient feature vector The gradient eigenvalue is large. When the cell belongs to the normal area, The gradient feature vector of the cell unit The smaller the initial adjustment value of the gradient eigenvalue.
[0049] Furthermore, the preliminary adjustment values of all gradient eigenvalues in the gradient eigenvector of each cell unit are sorted according to the order of the gradient eigenvalues in the gradient eigenvector to obtain the gradient adjustment vector of each cell unit.
[0050] Step S003: Obtain a gradient correction vector for each cell unit according to the difference between the gradient adjustment vectors of each cell unit and surrounding cell units.
[0051] It should be noted that due to the influence of environmental factors such as noise, only the gradient feature vector of a single cell unit is used for analysis, resulting in a reduced reliability of the initial adjustment value. In order to improve the enhancement effect of pelvic ultrasound images, it is necessary to adjust the gradient of the pixel points in combination with the gradient feature vectors of multiple cell units. For the normal area of the fallopian tube, due to the relatively straight edge of the normal area of the fallopian tube, each cell unit in the normal area is surrounded by cells with similar gradient feature vectors, while due to the local curvature and unevenness of the edge of the diseased area of the fallopian tube, the gradient feature vectors of each cell unit in the diseased area are not similar to those of the cell units around it, so the gradient feature vector is further adjusted accordingly.
[0052] Specifically, any cell unit is recorded as a target cell unit, and all cell units within the eight neighborhoods of the target cell unit are recorded as reference cell units of the target cell unit; it should be noted that if the number of cell units within the eight neighborhoods of the target cell unit is less than 8, only the existing cell units are recorded as reference cell units of the target cell unit; The calculation method of the gradient correction vector of a cell unit is:
[0053] ;
[0054] In the formula, For the The gradient correction vector of the cell unit; For the The gradient adjustment vector of each cell; For the The gradient adjustment vector of any reference cell of cells; is the minimum function; is a linear normalization function, and the normalization object is all cell units ; is the magnitude of the vector.
[0055] It should be noted that Indicates The smaller the value of the minimum difference between adjacent cells, the larger the value is. The greater the difference between the gradient adjustment vectors of a cell and its neighboring cells, the The greater the probability that the first cell belongs to the disease area, The greater the necessity of enhancing the gradient adjustment vector in each cell.
[0056] Step S004: obtaining the center position of the uterus according to the pelvic ultrasound image; obtaining the final feature vector of each cell unit according to the center position of the uterus and the gradient correction vector of each cell unit.
[0057] It should be noted that, since there are other tissue areas besides fallopian tubes in pelvic ultrasound images, other tissue areas may have similar features to the diseased areas of fallopian tubes, resulting in the enhancement of the gradients of some non-fallopian tube areas, which leads to a decrease in the enhancement quality of ultrasound images. Therefore, the position information of the cell unit is analyzed to further improve the accuracy of gradient enhancement. The blockage and swelling of the fallopian tubes generally occur at the end of the fallopian tubes, which are far away from the center of the uterus. Therefore, the gradient correction vector of the cell unit is adjusted according to the distance relationship between the cell unit and the center of the uterus. Since the pelvic image is symmetrical as a whole, the center of the uterus is determined accordingly.
[0058] Specifically, the pelvic ultrasound image is divided into two types of areas by the Otsu threshold method, and the area with the largest mean grayscale value of the pixels in the area is recorded as the main area; a two-dimensional rectangular coordinate system is established with the lower left corner of the pelvic ultrasound image as the origin, the horizontal direction as the x-axis, and the vertical direction as the y-axis; the mean of the x-coordinates corresponding to all the pixels in the main area is recorded as the horizontal coordinate of the uterus center; the mode of the y-coordinates corresponding to all the pixels in the main area is recorded as the vertical coordinate of the uterus center; the center position of the uterus is obtained according to the horizontal coordinate and the vertical coordinate of the uterus center; for any cell unit, the distance from the central pixel point of the cell unit to the center position of the uterus is recorded as the characteristic distance of the cell unit; the Otsu threshold method is a well-known technology, and the specific method is not introduced here; The calculation method of the final feature vector of a cell unit is:
[0059] ;
[0060] In the formula, For the The final feature vector of the cell unit; For the The characteristic distance of each cell; For the The gradient correction vector of the cell unit; It is a linear normalization function, and the normalization object is the characteristic distance of all cell units.
[0061] Step S005: Obtain the examination result of the infertility ultrasound image according to the final feature vectors of all cell units in the pelvic ultrasound image.
[0062] Specifically, the final feature vectors of all cell units in the pelvic ultrasound image are input into the SVW support vector machine to obtain the ultrasound imaging results of infertility, wherein the ultrasound imaging results of infertility include normal fallopian tube images or abnormal fallopian tube images; wherein, the training method of the SVW support vector machine is: obtaining a number of pelvic ultrasound images in the ultrasound department, obtaining the final feature vectors of all cell units in each pelvic ultrasound image according to the above method, and dividing all pelvic ultrasound images into two categories, namely, abnormal fallopian tube images and normal fallopian tube images, using the final feature vectors of all cell units in all pelvic ultrasound images and the classification results corresponding to all pelvic ultrasound images as data sets, training the SVW support vector machine, and outputting a normal fallopian tube image or an abnormal fallopian tube image. The specific training process is well known and will not be described in detail in this embodiment.
[0063] At this point, this embodiment is completed.
[0064] The present invention also provides an auxiliary analysis system for infertility ultrasound examination images, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor executes the computer program stored in the memory to implement the steps of the aforementioned auxiliary analysis method for infertility ultrasound examination images.
[0065] 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. An auxiliary analysis method for infertility ultrasound examination images, characterized in that: The method comprises the following steps: Acquire a pelvic ultrasound image; divide the pelvic ultrasound image into a number of cell units; obtain a gradient feature vector of each cell unit according to the gradient of the pixel points in each cell unit; the gradient feature vector includes a number of gradient feature values; According to the difference of the gradient eigenvalues in the gradient eigenvalue vector of each cell unit, a preliminary adjustment value of each gradient eigenvalue in the gradient eigenvalue vector of each cell unit is obtained; the preliminary adjustment values of all the gradient eigenvalues in the gradient eigenvalue vector of each cell unit are sorted to obtain the gradient adjustment vector of each cell unit; According to the difference between the gradient adjustment vectors of each cell unit and the surrounding cell units, the gradient correction vector of each cell unit is obtained; The center position of the uterus is obtained according to the pelvic ultrasound image; the final feature vector of each cell unit is obtained according to the center position of the uterus and the gradient correction vector of each cell unit; According to the final feature vectors of all cell units in the pelvic ultrasound image, the examination results of the infertility ultrasound image are obtained; Among them, according to the difference of the gradient eigenvalues in the gradient eigenvalue vector of each cell unit, the preliminary adjustment value of each gradient eigenvalue in the gradient eigenvalue vector of each cell unit is obtained, including: According to the difference of the gradient eigenvalues in the gradient eigenvector of each cell unit, the disease degree of the gradient eigenvector of each cell unit is obtained; According to the gradient eigenvalue in the gradient eigenvector of each cell unit, the necessity of maintaining each gradient eigenvalue in the gradient eigenvector of each cell unit is obtained; According to the disease degree of the gradient feature vector of each cell unit and each gradient feature value in the gradient feature vector of each cell unit and the necessity of maintaining the gradient feature value, a preliminary adjustment value of each gradient feature value in the gradient feature vector of each cell unit is obtained; According to the severity of the gradient feature vector of each cell unit and each gradient feature value in the gradient feature vector of each cell unit and the necessity of maintaining the gradient feature value, a preliminary adjustment value of each gradient feature value in the gradient feature vector of each cell unit is obtained, and the specific calculation method is as follows: ； In the formula, For the The gradient feature vector of the cell unit The initial adjustment value of the gradient eigenvalue; For the The sum of all gradient eigenvalues in the gradient eigenvector of a cell; For the The maximum gradient eigenvalue in the gradient eigenvector of the cell unit; is the number of bin directions; For the The first cell in the gradient feature vector Gradient eigenvalues; is the normalization function; For the The severity of the gradient feature vector of each cell; For the The gradient feature vector of the cell unit The necessity of maintaining the gradient eigenvalues.
2. The auxiliary analysis method of infertility ultrasound examination images according to claim 1, characterized in that: The step of obtaining the gradient feature vector of each cell unit according to the gradient of the pixel points in each cell unit includes the following specific steps: Divide 360 degrees into equal parts direction intervals, and each direction interval is recorded as a bin direction; among them, is the preset number of bin directions; get the value of each cell in A directional gradient histogram is obtained in each bin direction, wherein the directional gradient histogram contains the gradient eigenvalues in each bin direction, and all the gradient eigenvalues are sorted in the order of the bin direction corresponding to each gradient eigenvalue to obtain the gradient eigenvector of each cell unit.
3. The auxiliary analysis method of infertility ultrasound examination images according to claim 1, characterized in that: The step of sorting the preliminary adjustment values of all gradient feature values in the gradient feature vector of each cell unit to obtain the gradient adjustment vector of each cell unit includes the following specific steps: The preliminary adjustment values of all gradient eigenvalues in the gradient eigenvector of each cell unit are sorted according to the order of the gradient eigenvalues in the gradient eigenvector to obtain the gradient adjustment vector of each cell unit.
4. The auxiliary analysis method of infertility ultrasound examination images according to claim 1, characterized in that: The step of obtaining the gradient correction vector of each cell unit according to the difference between the gradient adjustment vectors of each cell unit and the surrounding cell units includes the following specific steps: Record any cell as the target cell, and record all cells within the eight neighborhoods of the target cell as reference cells of the target cell; calculate the The cell unit and the The modulus of the difference vector of the gradient adjustment vector of each reference cell unit of the cell unit, the normalized result of the smallest modulus and the The product of the gradient adjustment vectors of the cell units is recorded as The gradient correction vector of the cell unit.
5. The auxiliary analysis method of infertility ultrasound examination images according to claim 1, characterized in that: The method of obtaining the center position of the uterus according to the pelvic ultrasound image includes the following specific steps: The pelvic ultrasound image is divided into two types of areas by the Otsu threshold method, and the area with the largest mean grayscale value of the pixels in the area is recorded as the main area; a two-dimensional rectangular coordinate system is established with the lower left corner of the pelvic ultrasound image as the origin, the horizontal direction as the x-axis, and the vertical direction as the y-axis; the mean of the x-coordinates corresponding to all the pixels in the main area is recorded as the horizontal coordinate of the uterus center; the mode of the y-coordinates corresponding to all the pixels in the main area is recorded as the vertical coordinate of the uterus center; the center position of the uterus is obtained according to the horizontal coordinate and the vertical coordinate of the uterus center.
6. The auxiliary analysis method of infertility ultrasound examination images according to claim 1, characterized in that: The method of obtaining the final feature vector of each cell unit according to the center position of the uterus and the gradient correction vector of each cell unit includes the following specific steps: For any cell unit, the distance from the central pixel of the cell unit to the center of the uterus is recorded as the characteristic distance of the cell unit; The first The normalized result of the feature distance of the cell unit is the same as that of the The product of the gradient correction vectors of the cell units is recorded as The final feature vector of the cell.
7. The auxiliary analysis method of infertility ultrasound examination images according to claim 1, characterized in that: The method of obtaining the examination result of the infertility ultrasound image according to the final feature vectors of all cell units in the pelvic ultrasound image includes the following specific steps: Inputting the final feature vectors of all the cell units in the pelvic ultrasound image into the SVW support vector machine to obtain the ultrasound imaging examination results of infertility; the ultrasound imaging examination results of infertility include a normal fallopian tube image or an abnormal fallopian tube image; Among them, the training method of the SVW support vector machine is: obtain several pelvic ultrasound images in the ultrasound department, obtain the final feature vectors of all cell units in each pelvic ultrasound image, and divide all pelvic ultrasound images into two categories, namely, abnormal fallopian tube images and normal fallopian tube images. The final feature vectors of all cell units in all pelvic ultrasound images and the classification results corresponding to all pelvic ultrasound images are used as data sets to train the SVW support vector machine, and the output is a normal fallopian tube image or an abnormal fallopian tube image.
8. An auxiliary analysis system for infertility ultrasound examination images, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the computer program is executed by the processor, the steps of the auxiliary analysis method of infertility ultrasound examination images as described in any one of claims 1 to 7 are implemented.