Gastric cancer tissue typing auxiliary identification method based on digital pathological section

Abstract

The invention relates to the technical field of image processing, in particular to a stomach cancer tissue typing auxiliary recognition method based on digital pathological sections. According to the method, morphological feature recognition is carried out on edge information in a grayscale image of a slice, and normal cell edges and abnormal cell edges are screened out according to the size of morphological feature indexes and the morphology of a cell edge curve. A first abnormal area and a second abnormal area are screened out according to the proximity relation, then enhancement coefficients of the two abnormal areas are calculated and enhanced, and a gastric cancer tissue typing recognition result is obtained according to the enhanced stomach pathology image. According to the invention, the object of the enhancement process is only limited to the abnormal region, and different degrees of enhancement are realized for different types of abnormal regions, so that the calculation amount of the enhancement process is reduced, the enhancement process has pertinence, excessive enhancement cannot be caused, and a clear slice image is obtained for tissue typing identification; and the tissue typing recognition efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of image processing, and particularly relates to a method for assisting in the identification of gastric cancer tissue typing based on digital pathological sections. Background Art

[0002] In the process of using digital section technology to identify the typing of gastric cancer tissue, the image quality may be affected by various factors, such as errors in the section preparation process, uneven staining, insufficient resolution of imaging equipment, etc. These factors may cause the image to be blurred and details to be lost, thus affecting the identification effect of gastric cancer tissue typing. In the prior art, the section image is enhanced before the typing identification to improve the image quality and ensure the identification effect. However, the enhancement of the section image in the prior art is a global enhancement process. In fact, in addition to diseased cells, the image also includes normal cells. For the typing identification process, the information of diseased cells is more important than that of normal cells. If the section image is globally enhanced, the enhancement object will be non-targeted; and the contrast degree between diseased cells and normal cells in the image is not considered, resulting in excessive enhancement of the information of diseased cells, increasing the computational amount of the enhancement process and affecting the efficiency of tissue typing identification. Summary of the Invention

[0003] In order to solve the technical problem that when the prior art performs image enhancement on a section image, the setting of the enhancement degree is unreasonable, resulting in an increase in the computational amount of the enhancement process and affecting the efficiency of tissue typing identification, the object of the present invention is to provide a method for assisting in the identification of gastric cancer tissue typing based on digital pathological sections, and the specific technical solution adopted is as follows: The present invention provides a method for assisting in the identification of gastric cancer tissue typing based on digital pathological sections, and the method includes: Obtaining a grayscale image of a digital pathological section of the stomach; obtaining multiple cell edge curves in the grayscale image; For each cell edge curve, obtaining a morphological feature index of each cell edge curve according to the size and edge smoothness of the region enclosed by the cell edge curve; classifying the cell edge curves according to the morphological feature index to obtain multiple edge categories; screening out normal cell edges and abnormal cell edges according to the size and edge smoothness of the morphological feature index in each edge category; Taking the abnormal cell edges adjacent to the normal cell edges as the first abnormal cell edges, and other abnormal cell edges as the second abnormal cell edges; dividing multiple first abnormal regions according to the distances between the first abnormal cell edges; dividing multiple second abnormal regions according to the distances between the second abnormal cell edges; Obtain the enhancement coefficient for each first abnormal region based on the grayscale difference between the first abnormal region and the adjacent normal cell region, as well as the number of first abnormal cell edges within the first abnormal region, and perform enhancement; obtain the enhancement coefficient for each second abnormal region based on the number of second abnormal cell edges within the second abnormal region and perform enhancement; obtain the enhanced gastric pathological image; Obtain the recognition result of gastric cancer tissue typing based on the enhanced gastric pathological image.

[0004] Furthermore, the method for obtaining the cell edge curve includes: Obtain the edge pixel points in the grayscale image using an adaptive morphological edge detection algorithm; take each edge pixel point as a point to be grown. If there are other edge pixel points within the preset neighborhood range of the point to be grown, then take the other edge pixel points as new points to be grown and continue growing until growth is no longer possible, or the preset neighborhood range of the point to be grown contains the initial growth point. Take the growth result as a cell edge curve, and obtain all cell edge curves.

[0005] Furthermore, the method for obtaining the edge smoothness includes: Obtain the tangent angle difference between adjacent edge pixel points on the cell edge curve, and perform a negative correlation mapping on the average tangent angle difference to obtain the edge smoothness.

[0006] Furthermore, the method for obtaining the morphological feature index includes: Take the region enclosed by the cell edge curve as the cell region, take the maximum distance between edge pixel points on the cell edge curve as the size length of the cell region, take the product of the area and the size length of the cell region as the size scale feature of the cell region, and perform normalization processing on the product of the size scale feature and the edge smoothness to obtain the morphological feature index.

[0007] Furthermore, the screening of normal cell edges and abnormal cell edges includes: If the morphological feature index in the edge category is within the preset index range, then take the edge category as the edge category to be selected; if the average edge smoothness in the edge category to be selected is greater than the preset threshold, then take the edge category to be selected as the category composed of normal cell edges, and other edge categories are all categories composed of abnormal cell edges.

[0008] Furthermore, the screening method for the first abnormal cell edge includes: With each normal cell edge as the center, construct a neighborhood judgment region for each normal cell edge according to the preset neighborhood size, and the abnormal cell edges in the neighborhood judgment region are the first abnormal cell edges.

[0009] Further, the method for dividing the first abnormal region and the second abnormal region includes: Obtain the neighborhood judgment regions of the edges of each first abnormal cell, and take the edges of the first abnormal cells where the neighborhood judgment regions intersect as the edges of the same kind of first abnormal cells. The region composed of the edges of the same kind of first abnormal cells is used as the first abnormal region; Obtain the neighborhood judgment regions of the edges of each second abnormal cell, and take the edges of the second abnormal cells where the neighborhood judgment regions intersect as the edges of the same kind of second abnormal cells. The region composed of the edges of the same kind of second abnormal cells is used as the second abnormal region.

[0010] Further, the method for obtaining the enhancement coefficient of the first abnormal region includes: Take the region enclosed by the edges of the abnormal cells as the abnormal cell region, and the region enclosed by the edges of the normal cells as the normal cell region; for each edge of the first abnormal cell, obtain the gray-scale difference between the abnormal cell region enclosed by the edge of the first abnormal cell and the nearest normal cell region; obtain the average gray-scale difference in the first abnormal region, normalize the ratio of the number of edges of the first abnormal cells in the first abnormal region to the average gray-scale difference, obtain the initial enhancement coefficient of the first abnormal region, and take the sum value of the initial enhancement coefficient and the positive integer 1 as the enhancement coefficient of the first abnormal region.

[0011] Further, the method for obtaining the enhancement coefficient of the second abnormal region includes: Normalize the number of edges of the second abnormal cells in the second abnormal region to obtain the initial enhancement coefficient of the second abnormal region, and take the sum value of the initial enhancement coefficient and the positive integer 1 as the enhancement coefficient of the second abnormal region.

[0012] Further, take any one of the first abnormal region and the second abnormal region as the region to be enhanced; multiply the original pixel value of each pixel point in the region to be enhanced by the enhancement coefficient, obtain the enhanced pixel value and perform pixel value replacement to obtain the enhanced region.

[0013] The present invention has the following beneficial effects: The present invention first performs morphological feature recognition on the edge information in the grayscale image of the slice, and screens out the normal cell edges and abnormal cell edges according to the magnitudes of the morphological feature indexes and the morphology of the cell edge curves. Since there are obvious morphological differences between normal cells and abnormal cells, which are reflected in size changes and edge morphologies, the normal cell edges and abnormal cell edges can be effectively screened out by using the morphological feature indexes and the edge smoothness. The first abnormal region and the second abnormal region are screened out according to the adjacency relationship, and then enhancement coefficients are obtained and enhanced for the two abnormal regions respectively, so that the object of the enhancement process is limited to the abnormal regions, and different degrees of enhancement are performed for different types of abnormal regions, reducing the computational amount in the enhancement process, making the enhancement process targeted and not causing over-enhancement, obtaining a clear slice image for the recognition of tissue typing, and improving the efficiency of tissue typing recognition. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present invention, and those of ordinary skill in the art can obtain other drawings based on these drawings without creative efforts.

[0015] Figure 1 It is a flowchart of a method for assisting in the recognition of gastric cancer tissue typing based on digital pathological slices provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0016] In order to further elaborate on the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following will, in conjunction with the drawings and preferred embodiments, describe in detail the specific implementation manner, structure, features and effects of a method for assisting in the recognition of gastric cancer tissue typing based on digital pathological slices proposed by the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures or characteristics in one or more embodiments can be combined in any suitable form.

[0017] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the technical field to which the present invention belongs.

[0018] The following will specifically describe the specific solution of a method for assisting in the recognition of gastric cancer tissue typing based on digital pathological slices provided by the present invention with reference to the drawings.

[0019] Please refer to Figure 1, which shows a flowchart of a method for assisting in the identification of gastric cancer tissue typing based on digital pathological sections provided by an embodiment of the present invention. The method includes: Step S1: Obtain a grayscale image of a gastric digital pathological section; obtain multiple cell edge curves in the grayscale image.

[0020] The embodiment of the present invention aims to perform targeted enhancement on the section image of gastric cancer tissue before tissue typing identification. Therefore, it is necessary to obtain the image of the gastric digital pathological section and convert it into a grayscale image for subsequent morphological analysis. In the embodiment of the present invention, the TDI-CCD linear scanning technology is used to quickly and comprehensively convert the information on the glass slide into data slice information.

[0021] The grayscale image contains cell information of the stomach, including normal cells and abnormal cells caused by gastric cancer. In order to perform targeted enhancement on the image, it is first necessary to distinguish between diseased cells and normal cells. There are obvious morphological differences between diseased cells and normal cells. Therefore, it is first necessary to extract the cell edge curves in the grayscale image, and then use morphological features to identify them in subsequent steps.

[0022] Preferably, in the embodiment of the present invention, an adaptive morphological edge detection algorithm is used to obtain edge pixel points in the grayscale image; each edge pixel point is used as a point to be grown. If there are other edge pixel points within the preset neighborhood range of the point to be grown, then the other edge pixel points are used as new points to be grown and continue to grow until growth is no longer possible, or the preset neighborhood range of the point to be grown contains the initial growth point. The growth result is used as a cell edge curve, and all cell edge curves are obtained.

[0023] In an embodiment of the present invention, the preset neighborhood range is set to an 8-neighborhood range. Further considering that other edge information in the image besides cell edges is likely to cause interference, in an embodiment of the present invention, after obtaining the growth result, according to the length of the edge corresponding to the production result, the growth results with smaller lengths are removed. Specifically, the growth results with smaller lengths can be screened out by setting thresholds, clustering classification, etc., which are all well-known technical means in the art and will not be elaborated and limited specifically.

[0024] Step S2: For each cell edge curve, obtain the morphological feature index of each cell edge curve according to the size of the area enclosed by the cell edge curve and the edge smoothness; classify the cell edge curves according to the morphological feature index to obtain multiple edge categories; screen out normal cell edges and abnormal cell edges according to the size of the morphological feature index and the edge smoothness in each edge category.

[0025] For normal cells, their edges are relatively regular, the boundary between cells is clear, showing smooth and continuous edge features. However, due to lesions, the edges of abnormal cells will show irregular shapes such as serrated, wavy or broken. And for normal cells, their morphological sizes are relatively consistent, and the sizes of cell regions are relatively similar. Therefore, based on the size of the cell region and the smoothness of the edge, the attributes of the cell edge curve can be classified. Therefore, in the embodiments of the present invention, the morphological feature indexes of each cell edge curve are first obtained according to the size of the region enclosed by the cell edge curve and the smoothness of the edge. That is, the morphological feature indexes can simultaneously characterize the features of cell size and edge smoothness. Further, the cell edge curves can be classified according to the morphological feature indexes to obtain multiple edge categories, that is, the cell edge curves in one edge category are regarded as the edges of the same kind of cells. According to the size of the morphological feature indexes of the edge categories and the smoothness of the edges, the normal cell edges and abnormal cell edges can be screened out.

[0026] Preferably, in an embodiment of the present invention, the method for obtaining the edge smoothness includes: Obtain the tangent angle difference between adjacent edge pixel points on the cell edge curve, perform a negative correlation mapping on the average tangent angle difference to obtain the edge smoothness. It should be noted that since the cell edge cannot be a straight line, the average tangent angle difference will not be 0. Therefore, in an embodiment of the present invention, the reciprocal of the average tangent angle difference is used as the edge smoothness.

[0027] Preferably, in an embodiment of the present invention, the method for obtaining the morphological feature indexes includes: Take the region enclosed by the cell edge curve as the cell region, take the maximum distance between edge pixel points on the cell edge curve as the size length of the cell region, take the product of the area of the cell region and the size length as the size scale feature of the cell region, and perform a normalization process on the product of the size scale feature and the edge smoothness to obtain the morphological feature indexes. That is, the larger the size scale feature and the larger the edge smoothness, the larger the morphological feature indexes.

[0028] It should be noted that the cell edge curve may not be a closed edge and cannot enclose a closed region. Therefore, the method of curve fitting needs to be used to complete the edge to obtain the cell region. Curve fitting can adopt fitting methods such as the least squares method, which are well-known technical means in the art and will not be elaborated here.

[0029] It should be noted that the normalization process can be implemented by other basic mathematical means such as range normalization and function mapping method, which are all well-known technical means in the art and will not be elaborated and limited here.

[0030] In the embodiments of the present invention, in order to centrally analyze the cell edge curve and further reduce the amount of calculation, the cell edge curve is classified according to morphological feature indicators to obtain multiple edge categories. In the embodiments of the present invention, considering that the morphological feature indicator is a normalized value with a value range between 0 and 1, the morphological feature indicator is divided into ten intervals with 0.1 as an interval, and each interval corresponds to an edge category, thus completing the classification of edge categories.

[0031] Preferably, in an embodiment of the present invention, screening out normal cell edges and abnormal cell edges includes: If the morphological feature indicator in the edge category is within a preset indicator interval, then the edge category is used as a candidate edge category; if the average edge smoothness in the candidate edge category is greater than a preset threshold, then the candidate edge category is used as the category composed of normal cell edges, and other edge categories are all categories composed of abnormal cell edges.

[0032] It should be noted that the embodiments of the present invention are directed to image enhancement that simultaneously includes normal cells and abnormal cells, so there is no situation where the image does not contain normal cells or does not contain abnormal cells. Therefore, the normal cell edge category and the abnormal cell edge category can be accurately identified through the setting of two conditions. The preset indicator interval and the preset threshold can be specifically set according to the parameters of the camera and the actual type of cells. For example, if the cells in the field of view are relatively small due to the camera parameters, the indicator interval should be set smaller, so no further limitation and elaboration are provided here.

[0033] Step S3: Regarding the abnormal cell edges adjacent to the normal cell edges as the first abnormal cell edges, and other abnormal cell edges as the second abnormal cell edges; dividing multiple first abnormal regions according to the distances between the first abnormal cell edges; dividing multiple second abnormal regions according to the distances between the second abnormal cell edges.

[0034] In the embodiments of the present invention, considering that the abnormal cells adjacent to the normal cells already have a relatively obvious contrast, their image features can be highlighted without excessive enhancement; while for the abnormal cells far from the normal cells, they can be enhanced normally. Therefore, the abnormal cells in the two states need to be analyzed separately, so it is necessary to classify them. Regarding the abnormal cell edges adjacent to the normal cell edges as the first abnormal cell edges, and other abnormal cell edges as the second abnormal cell edges.

[0035] For abnormal cells, since they are caused by canceration, their morphology is relatively complex. If there is a state of abnormal cell aggregation, it will lead to morphological overlap, resulting in inaccurate information and affecting subsequent tissue typing recognition. Therefore, when performing image enhancement later, it is necessary to consider the aggregation state of abnormal cells, that is, the more aggregated, the greater the degree of image enhancement required. Therefore, multiple first abnormal regions can be divided according to the distance between the edges of the first abnormal cells; multiple second abnormal regions can be divided according to the distance between the edges of the second abnormal cells.

[0036] Preferably, in an embodiment of the present invention, the screening method for the edges of the first abnormal cells includes: Taking each normal cell edge as the center, constructing a neighborhood judgment region for each normal cell edge according to a preset neighborhood size, and the abnormal cell edges in the neighborhood judgment region are used as the edges of the first abnormal cells. In an embodiment of the present invention, the preset neighborhood size can be set to the size length of the central cell region, and it is considered that if the neighborhood judgment region intersects with the abnormal cell edge, the judgment condition for the edge of the first abnormal cell is satisfied.

[0037] Furthermore, the method for dividing the first abnormal region and the second abnormal region includes: Obtaining the neighborhood judgment region of each edge of the first abnormal cell, taking the edges of the first abnormal cells where the neighborhood judgment regions intersect as the same-kind edges of the first abnormal cells, and the region composed of the same-kind edges of the first abnormal cells is used as the first abnormal region.

[0038] Obtaining the neighborhood judgment region of each edge of the second abnormal cell, taking the edges of the second abnormal cells where the neighborhood judgment regions intersect as the same-kind edges of the second abnormal cells, and the region composed of the same-kind edges of the second abnormal cells is used as the second abnormal region.

[0039] Step S4: Obtain the enhancement coefficient of each first abnormal region according to the gray-scale difference between the first abnormal region and the adjacent normal cell edges, and the number of edges of the first abnormal cells in the first abnormal region, and perform enhancement; obtain the enhancement coefficient of each second abnormal region according to the number of edges of the second abnormal cells in the second abnormal region and perform enhancement; obtain the enhanced gastric pathological image.

[0040] As described in step S3, for the first abnormal region, when enhancing it, the neighboring normal cells need to be considered. The greater the gray-scale difference between the first abnormal region and the neighboring normal cell region, the less significant the enhancement required for the first abnormal region, as it already has an obvious contrast with the normal cells. The first abnormal region is similar to the second abnormal region in that the number of abnormal cells represented by the edges of the abnormal cells within the region needs to be considered for both. That is, the more the number of edges of the abnormal cells, the more concentrated the abnormal cells in the abnormal region, and the greater the enhancement degree required. Therefore, based on the gray-scale difference between the first abnormal region and the neighboring normal cell region, and the number of the first abnormal cell edges within the first abnormal region, the enhancement coefficient of each first abnormal region is obtained and enhanced; based on the number of the second abnormal cell edges within the second abnormal region, the enhancement coefficient of each second abnormal region is obtained and enhanced; and then the enhanced gastric pathological image is obtained.

[0041] Preferably, in an embodiment of the present invention, the method for obtaining the enhancement coefficient of the first abnormal region includes: The region enclosed by the edges of the abnormal cells is taken as the abnormal cell region, and the region enclosed by the edges of the normal cells is taken as the normal cell region; for each first abnormal cell edge, the gray-scale difference between the abnormal cell region enclosed by the first abnormal cell edge and the nearest normal cell region is obtained. It should be noted that the nearest means the normal cell region with the shortest distance, and the distance between cell regions can be characterized by the distance between the centroid points of the regions, and the gray-scale difference is the absolute value of the difference between the average gray-scale values within the regions.

[0042] Each abnormal cell region in the first abnormal region will correspond to a gray-scale difference. Therefore, the average gray-scale difference in the first abnormal region is obtained, and the ratio of the number of the first abnormal cell edges in the first abnormal region to the average gray-scale difference is normalized to obtain the initial enhancement coefficient of the first abnormal region. Since the initial enhancement coefficient is a value between 0 and 1, in the embodiment of the present invention, the sum of the initial enhancement coefficient and the positive integer 1 is used as the enhancement coefficient of the first abnormal region, that is, the enhancement coefficient is an enhancement multiple, and its value range is from 1 to 2.

[0043] Preferably, in an embodiment of the present invention, the method for obtaining the enhancement coefficient of the second abnormal region includes: The number of the second abnormal cell edges within the second abnormal region is normalized to obtain the initial enhancement coefficient of the second abnormal region, and the sum of the initial enhancement coefficient and the positive integer 1 is used as the enhancement coefficient of the second abnormal region.

[0044] Preferably, in an embodiment of the present invention, a linear enhancement method is used to enhance the pixel values in the abnormal region. Specifically: any one of the first abnormal region and the second abnormal region is used as the region to be enhanced; the original pixel value of each pixel point in the region to be enhanced is multiplied by the enhancement coefficient to obtain the enhanced pixel value and perform pixel value replacement to obtain the enhanced region.

[0045] Step S5: Obtain the recognition result of gastric cancer tissue typing based on the enhanced gastric pathological image.

[0046] In the embodiment of the present invention, the recognition of gastric cancer tissue typing is based on a trained recognition model, which can be regarded as the prior art and will not be elaborated here.

[0047] In summary, the embodiment of the present invention performs morphological feature recognition on the edge information in the grayscale image of the slice, and screens out the normal cell edges and abnormal cell edges according to the size of the morphological feature index and the morphology of the cell edge curve. The first abnormal region and the second abnormal region are screened out according to the adjacency relationship, and then the enhancement coefficients are obtained and enhanced for the two abnormal regions respectively. The recognition result of gastric cancer tissue typing is obtained based on the enhanced gastric pathological image. The object of the enhancement process of the present invention is limited to the abnormal region, and different degrees of enhancement are performed for different types of abnormal regions, reducing the computational amount of the enhancement process, making the enhancement process targeted, and not causing over-enhancement, obtaining a clear slice image for the recognition of tissue typing, and improving the efficiency of tissue typing recognition.

[0048] It should be noted that: the above sequence of the embodiments of the present invention is only for description and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0049] Each embodiment in this specification is described in a progressive manner. The same or similar parts among the embodiments can be referred to each other, and the key points of each embodiment are the differences from other embodiments.

Claims

1. An auxiliary recognition method for gastric cancer tissue typing based on digital pathological sections, characterized in that, The method includes: Obtaining a grayscale image of a gastric digital pathological section; obtaining multiple cell edge curves in the grayscale image; For each cell edge curve, obtaining a morphological feature index of each cell edge curve according to the size and edge smoothness of the region enclosed by the cell edge curve; classifying the cell edge curves according to the morphological feature index to obtain multiple edge categories; screening out normal cell edges and abnormal cell edges according to the size of the morphological feature index and the edge smoothness in each edge category; Taking the abnormal cell edges adjacent to the normal cell edges as the first abnormal cell edges, and other abnormal cell edges as the second abnormal cell edges; dividing multiple first abnormal regions according to the distances between the first abnormal cell edges; dividing multiple second abnormal regions according to the distances between the second abnormal cell edges; Obtaining an enhancement coefficient for each first abnormal region according to the gray difference between the first abnormal region and the adjacent normal cell region, and the number of first abnormal cell edges in the first abnormal region, and performing enhancement; obtaining an enhancement coefficient for each second abnormal region according to the number of second abnormal cell edges in the second abnormal region, and performing enhancement; obtaining an enhanced gastric pathological image; Obtaining an identification result of gastric cancer tissue typing according to the enhanced gastric pathological image.

2. The gastric cancer tissue typing auxiliary recognition method based on digital pathological sections according to claim 1, wherein The method for obtaining the cell edge curve includes: Obtaining edge pixel points in the grayscale image by using an adaptive morphological edge detection algorithm; taking each edge pixel point as a point to be grown. If there are other edge pixel points within the preset neighborhood range of the point to be grown, taking the other edge pixel points as new points to be grown and continuing to grow until growth is no longer possible, or the preset neighborhood range of the point to be grown contains the initial growth point, taking the growth result as a cell edge curve, and obtaining all cell edge curves.

3. The gastric cancer tissue typing auxiliary recognition method based on digital pathological sections according to claim 1, wherein The method for obtaining the edge smoothness includes: Obtaining the tangent angle difference between adjacent edge pixel points on the cell edge curve, and performing a negative correlation mapping on the average tangent angle difference to obtain the edge smoothness.

4. A method for assisting in the identification of gastric cancer tissue typing based on digital pathological sections according to claim 1, characterized in that The method for obtaining the morphological feature index includes: Taking the region enclosed by the cell edge curve as the cell region, taking the maximum distance between edge pixel points on the cell edge curve as the size length of the cell region, taking the product of the area and the size length of the cell region as the size scale feature of the cell region, and performing normalization processing on the product of the size scale feature and the edge smoothness to obtain the morphological feature index.

5. A method for assisting in the identification of gastric cancer tissue typing based on digital pathological sections according to claim 1, characterized in that, The screening out of normal cell edges and abnormal cell edges includes: If the morphological feature index in the edge category is within a preset index interval, taking the edge category as a candidate edge category; if the average edge smoothness in the candidate edge category is greater than a preset threshold, taking the candidate edge category as the category composed of normal cell edges, and other edge categories are all categories composed of abnormal cell edges.

6. The gastric cancer tissue typing auxiliary recognition method based on digital pathological sections according to claim 1, wherein The screening method for the first abnormal cell edges includes: Centered on the edge of each normal cell, a neighborhood judgment area for the edge of each normal cell is constructed according to a preset neighborhood size, and the abnormal cell edge in the neighborhood judgment area is used as the first abnormal cell edge.

7. A method for assisting in the identification of gastric cancer tissue typing based on digital pathological sections according to claim 6, characterized in that, The method for dividing the first abnormal area and the second abnormal area includes: Obtain the neighborhood judgment area of each first abnormal cell edge, and use the first abnormal cell edges where the neighborhood judgment areas intersect as the same type of first abnormal cell edges. The area composed of the same type of first abnormal cell edges is used as the first abnormal area; Obtain the neighborhood judgment area of each second abnormal cell edge, and use the second abnormal cell edges where the neighborhood judgment areas intersect as the same type of second abnormal cell edges. The area composed of the same type of second abnormal cell edges is used as the second abnormal area.

8. The method for assisting in the identification of gastric cancer tissue typing based on digital pathological sections according to claim 1, wherein The method for obtaining the enhancement coefficient of the first abnormal area includes: Use the area enclosed by the abnormal cell edges as the abnormal cell area, and the area enclosed by the normal cell edges as the normal cell area; for each first abnormal cell edge, obtain the gray level difference between the abnormal cell area enclosed by the first abnormal cell edge and the nearest normal cell area; obtain the average gray level difference in the first abnormal area, normalize the ratio of the number of first abnormal cell edges in the first abnormal area to the average gray level difference to obtain the initial enhancement coefficient of the first abnormal area, and use the sum of the initial enhancement coefficient and the positive integer 1 as the enhancement coefficient of the first abnormal area.

9. A method for assisting in the identification of gastric cancer tissue typing based on digital pathological sections according to claim 1, characterized in that The method for obtaining the enhancement coefficient of the second abnormal area includes: Normalize the number of second abnormal cell edges in the second abnormal area to obtain the initial enhancement coefficient of the second abnormal area, and use the sum of the initial enhancement coefficient and the positive integer 1 as the enhancement coefficient of the second abnormal area.

10. A method for assisting in the identification of gastric cancer tissue typing based on digital pathological sections according to claim 1, wherein, Use either the first abnormal area or the second abnormal area as the area to be enhanced; Multiply the original pixel value of each pixel point in the area to be enhanced by the enhancement coefficient, obtain the enhanced pixel value and perform pixel value replacement to obtain the enhanced area.

Images