A pathological section image analysis method and system

Abstract

The present invention relates to the field of computer vision technology, and specifically to a pathological slice image analysis method and system, comprising the following steps: based on pathological slice image information, combined with image preprocessing, edge information in the image is detected and extracted, color information is analyzed, color distribution characteristics in the image are evaluated, and image edge and color information is generated. In the present invention, by analyzing edge and color information, combined with the extraction of texture features, the identification process of tissue structure is optimized, the identification and segmentation of lesion tissue is made more accurate, the identification and segmentation of lesion tissue is performed using tissue structure feature data, the location of the lesion area and the identification of the type are achieved, and the visual presentation quality of the data and the intuitiveness of the diagnosis are improved by mapping the pathological slice image into a three-dimensional model and annotating it, helping medical professionals to fully understand the pathological conditions in the slice image and make accurate treatment decisions.

Description

Technical Field

[0001] The present invention relates to the field of computer vision technology, and in particular to a pathological slice image analysis method and system. Background Art

[0002] The field of computer vision technology focuses on enabling computers to interpret information from images or multi-dimensional data. By simulating the mechanisms of the human visual system, computers can recognize and process objects, scenes, and various visual elements, and make judgments or responses based on the information. It involves the comprehensive application of multiple technologies such as image processing, machine learning, pattern recognition, and artificial intelligence, so that machines can realize the visual perception function of the human eye. It is applied to environmental perception of autonomous vehicles, quality control in industrial automation, security monitoring, interactive media, automated retail, and medical image analysis.

[0003] Among them, a pathological section image analysis method aims to use computer vision technology to automatically analyze pathological section images in the medical field. By automatically identifying and quantifying cells, tissue structures and lesion characteristics in pathological images, it assists doctors in making accurate assessments in diagnosis, treatment decisions and pathological research, automatically detects and quantifies the characteristics of lesion tissues, provides key diagnostic information, improves the accuracy and efficiency of diagnosis, reduces the workload of medical professionals when processing batch samples, and improves the efficiency and accuracy of pathological diagnosis.

[0004] Traditional pathology slice image analysis methods have poor accuracy and efficiency when processing complex pathology slice images. They lack the ability to recognize specific pathological features in image preprocessing and feature extraction. They cannot effectively distinguish subtle differences in the integration of texture and color information, resulting in the loss of diagnostic information. For the segmentation of lesion areas and the analysis of lesion types, they cannot comprehensively utilize multidimensional data for effective analysis, which limits the accuracy of diagnosis and affects the treatment effect of patients. When faced with batches of complex data, the processing speed is slow and inefficient, which increases the workload of medical professionals. Summary of the invention

[0005] In order to solve the technical problem of poor processing efficiency in the prior art, an embodiment of the present invention provides a pathological slice image analysis method and system. The technical solution is as follows:

[0006] In one aspect, a pathological section image analysis method is provided, the method comprising:

[0007] S1: Based on the pathological slice image information and combined with image preprocessing, the edge information in the image is detected and extracted, the color information is analyzed, the color distribution characteristics in the image are evaluated, and the image edge and color information is generated;

[0008] S2: Based on the edge and color information of the image, extract the texture features of multiple positions in the image, including the directionality, roughness and contrast of the texture, combine the edge information, analyze and extract the shape features in the image, identify the tissue structure and record the type information, and generate tissue structure feature data;

[0009] S3: using the tissue structure feature data, identifying the diseased tissue by comparing the differences in texture and color at multiple locations in the tissue structure region, and segmenting the image to generate a segmented image of the lesion location;

[0010] S4: segmenting the image based on the lesion location, analyzing the lesion tissue structure, calculating geometric properties, including area, perimeter, and complexity, analyzing the lesion type by comparing the data features with data of multiple known lesion types, and generating a lesion type analysis result;

[0011] S5: According to the lesion type analysis result and the type information of the tissue structure, a three-dimensional model of the target tissue is established and the pathological slice image is mapped into the model, and various information is annotated to generate a three-dimensional pathological tissue model.

[0012] As a further solution of the present invention, the image edge and color information includes an edge information data set, color extraction information, and image preprocessing parameters; the tissue structure feature data includes a texture information data set, a shape analysis result, and a tissue structure type label; the lesion position segmentation image includes a lesion area boundary, texture feature mapping data within the lesion area, and a lesion area color distribution map; the lesion type analysis result includes a lesion type recognition result, lesion area geometric properties, and a matching degree calculation result; and the three-dimensional pathological tissue model includes a three-dimensional model construction result, annotation information, and a slice image mapping result.

[0013] As a further solution of the present invention, based on the pathological slice image information, combined with image preprocessing, the edge information in the image is detected and extracted, the color information is analyzed, the color distribution characteristics in the image are evaluated, and the steps of generating the image edge and color information are specifically as follows:

[0014] S101: Preprocessing the image based on the pathological slice image information, optimizing the visibility of edge information in the image by adjusting the contrast and brightness of the image, and generating an image preprocessing record;

[0015] S102: Detecting and extracting edge information in the pathological slice image based on the image preprocessing record, and generating an edge information extraction record;

[0016] S103: extracting the color information and evaluating the color distribution characteristics in the image based on the edge information extraction record, and generating the image edge and color information.

[0017] As a further solution of the present invention, the specific formula for detecting and extracting edge information in the pathological slice image is:

[0018]

[0019] Among them, G represents the edge strength value, G x Indicates the horizontal gradient of each pixel in the image, G y It represents the vertical gradient size of each pixel in the image. x represents the horizontal coordinate in the image, which is used to calibrate the position of the pixel in the image matrix. y represents the vertical coordinate in the image, which is used to calibrate the position of the pixel in the image matrix. c is the adjustment coefficient, which is used to adjust the weight ratio of gradient synthesis to the average value.

[0020] As a further solution of the present invention, based on the edge and color information of the image, texture features at multiple positions in the image are extracted, including the directionality, roughness and contrast of the texture, and the shape features in the image are analyzed and extracted in combination with the edge information, the tissue structure is identified and the type information is recorded, and the steps of generating tissue structure feature data are specifically as follows:

[0021] S201: Analyzing texture information of the pathological slice image using the image edge and color information, including measuring the directionality and roughness of textures at multiple locations, and generating texture analysis results;

[0022] S202: Based on the texture analysis result and in combination with edge information, shape features in the image are analyzed and extracted, tissue structures are identified and marked, and a structure recognition map is generated;

[0023] S203: Based on the structure recognition graph, the shape of the tissue structure is analyzed to identify the type of the tissue structure, and tissue structure feature data is generated by combining the shape features and texture features of the tissue.

[0024] As a further solution of the present invention, the tissue structure feature data is used to identify the diseased tissue by comparing the differences in texture and color at multiple locations in the tissue structure area, and the image is segmented. The steps of generating a segmented image of the lesion location are specifically as follows:

[0025] S301: Analyze the color and texture parameters of the tissue area based on the tissue structure feature data, identify the lesion area by comparing and analyzing the color and texture information of multiple positions, and generate lesion calibration data;

[0026] S302: Segmenting the lesion region based on the lesion calibration data, separating the normal tissue and the lesion tissue region, and generating lesion region segmentation data;

[0027] S303: Based on the lesion region segmentation data, the segmentation boundary is adjusted to optimize the visibility of the lesion edge and the integrity of the lesion region, and a lesion position segmentation image is generated.

[0028] As a further solution of the present invention, the steps of segmenting the image based on the lesion position, analyzing the lesion tissue structure, calculating geometric properties, including area, perimeter and complexity, analyzing the lesion type by comparing the data features with data of multiple known lesion types, and generating the lesion type analysis result are specifically as follows:

[0029] S401: Based on the lesion position segmentation image, the area and perimeter of the target lesion area are measured to generate geometric attribute information;

[0030] S402: Based on the geometric attribute information, by measuring the irregularity of the boundary of the lesion area and the length of the broken line, the shape complexity index of the lesion area is calculated to generate a complexity analysis result;

[0031] S403: Based on the complexity analysis result, by comparing with geometric features of multiple known lesion types, the lesion type in the target slice image is identified to generate a lesion type analysis result.

[0032] As a further solution of the present invention, the specific formula for identifying the lesion type in the target slice image is:

[0033]

[0034] Among them, d represents the distance from the support vector to the decision boundary, n represents the total number of features involved in the calculation, i represents the index of the current feature, and w i represents the weight of the i-th feature, x i represents the value of the i-th feature, and b represents the bias term of the decision boundary.

[0035] As a further solution of the present invention, according to the lesion type analysis result and the type information of the tissue structure, a three-dimensional model of the target tissue is established and the pathological slice image is mapped into the model, and various information is annotated. The steps of generating the three-dimensional pathological tissue model are specifically as follows:

[0036] S501: Based on the lesion type analysis result and according to the type information of the target tissue structure, a three-dimensional model of the target tissue is constructed to obtain a model construction record;

[0037] S502: Building a record based on the model, mapping the pathological slice image into the model, and aligning the texture and color to obtain an image information mapping record;

[0038] S503: Based on the image information mapping record, annotation information is added to the model, including edge information, lesion area, and lesion type, to obtain a three-dimensional pathological tissue model.

[0039] On the other hand, a pathological section image analysis system is provided, which is applied to the pathological section image analysis method, and the system comprises:

[0040] The slice image processing module adjusts the brightness and contrast of the image based on the pathological slice image information, combines noise removal, and generates image preprocessing results;

[0041] The edge detection and color analysis module extracts edge information, analyzes the color distribution of the image, and generates image edge and color information based on the image preprocessing result;

[0042] The texture and shape analysis module analyzes the texture features in the image, including directionality and roughness, based on the image edge and color information, and extracts shape features, combines the edge detection results, identifies the tissue structure, and generates tissue structure feature data;

[0043] The lesion region recognition module uses the tissue structure feature data to identify the lesion region by comparing the texture and color information of multiple positions, and performs image segmentation on the lesion region to generate a lesion position segmentation image;

[0044] The lesion image analysis module segments the image based on the lesion location and calculates the geometric properties of the lesion area, including area, perimeter, and complexity. It analyzes the lesion type by comparing it with data of multiple known lesion types and generates lesion type analysis results.

[0045] The three-dimensional modeling and mapping module constructs a three-dimensional model of the target tissue according to the lesion type analysis results and the type information of the tissue in the slice image, maps the pathological slice image data to the model, performs information annotation, and generates a three-dimensional pathological tissue model.

[0046] The beneficial effects brought about by the technical solution provided by the embodiment of the present invention include at least:

[0047] By analyzing edge and color information and extracting texture features, the tissue structure recognition process is optimized, making the identification and segmentation of diseased tissue more accurate. The tissue structure feature data is used to identify and segment diseased tissue, and the location and type of diseased areas can be identified. By mapping pathological slice images into three-dimensional models and annotating them, the visual presentation quality of data and the intuitiveness of diagnosis are improved, helping medical professionals to fully understand the pathological conditions in the slice images and make accurate treatment decisions. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. 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 creative work.

[0049] Figure 1 It is a schematic diagram of the workflow of the present invention;

[0050] Figure 2 This is a detailed flow chart of S1 of the present invention;

[0051] Figure 3 This is a detailed flow chart of S2 of the present invention;

[0052] Figure 4 This is a detailed flow chart of S3 of the present invention;

[0053] Figure 5 This is a detailed flow chart of S4 of the present invention;

[0054] Figure 6 This is a detailed flow chart of S5 of the present invention;

[0055] Figure 7 It is a system flow chart of the present invention. DETAILED DESCRIPTION

[0056] The technical solution of the present invention is described below in conjunction with the accompanying drawings.

[0057] In the embodiments of the present invention, words such as "exemplarily" and "for example" are used to indicate examples, illustrations or explanations. Any embodiment or design described as "example" in the present invention should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of the word "example" is intended to present the concept in a specific way. In addition, in the embodiments of the present invention, the meaning expressed by "and / or" can be both, or it can be either of the two.

[0058] In the embodiments of the present invention, "image" and "picture" can sometimes be used interchangeably. It should be noted that when the difference between them is not emphasized, the meanings they intend to express are the same. "of", "corresponding, relevant" and "corresponding" can sometimes be used interchangeably. It should be noted that when the difference between them is not emphasized, the meanings they intend to express are the same.

[0059] In the embodiments of the present invention, sometimes a subscript such as W1 may be written as a non-subscript such as W1. When the difference is not emphasized, the meanings to be expressed are the same.

[0060] In order to make the technical problems, technical solutions and advantages to be solved by the present invention more clear, a detailed description will be given below with reference to the accompanying drawings and specific embodiments.

[0061] The embodiment of the present invention provides a pathological section image analysis method, such as Figure 1 The process flow chart of the pathological slice image analysis method shown in the figure may include the following steps:

[0062] S1: Based on the pathological slice image information and combined with image preprocessing, the edge information in the image is detected and extracted, the color information is analyzed, the color distribution characteristics in the image are evaluated, and the image edge and color information is generated;

[0063] S2: Based on the image edge and color information, the texture features of multiple locations in the image are extracted, including the texture directionality, roughness and contrast. Combined with the edge information, the shape features in the image are analyzed and extracted, the tissue structure is identified and the type information is recorded to generate tissue structure feature data;

[0064] S3: Using tissue structure feature data, by comparing the differences in texture and color at multiple locations in the tissue structure area, the lesion tissue is identified, and the image is segmented to generate a segmented image of the lesion location;

[0065] S4: Segment the image based on the lesion location, analyze the lesion tissue structure, calculate the geometric properties, including area, perimeter and complexity, analyze the lesion type by comparing the data features with the data of multiple known lesion types, and generate the lesion type analysis results;

[0066] S5: According to the lesion type analysis results and the type information of the tissue structure, a three-dimensional model of the target tissue is established and the pathological slice image is mapped into the model, and various information is annotated to generate a three-dimensional pathological tissue model.

[0067] Image edge and color information includes edge information data set, color extraction information, and image preprocessing parameters. Tissue structure feature data includes texture information data set, shape analysis results, and tissue structure type labels. Lesion location segmentation image includes lesion area boundary, texture feature mapping data within the lesion area, and lesion area color distribution map. Lesion type analysis results include lesion type recognition results, lesion area geometric properties, and matching degree calculation results. Three-dimensional pathological tissue model includes three-dimensional model construction results, annotation information, and slice image mapping results.

[0068] See also Figure 2 ,Based on the pathological slice image information, combined with image preprocessing, the edge information in the image is detected and extracted, the color information is analyzed, and the color distribution characteristics in the image are evaluated. The specific steps for generating image edge and color information are as follows:

[0069] S101: Preprocessing the image based on the pathological slice image information, optimizing the visibility of edge information in the image by adjusting the contrast and brightness of the image, and generating an image preprocessing record;

[0070] In sub-step S101, the pathological slice image information is read, the original image data is loaded into the memory, the histogram equalization technology is applied for preprocessing, the grayscale histogram of the image is analyzed, and the original histogram is converted into a uniformly distributed histogram. The conversion is achieved by calculating the cumulative distribution function. Each value of the cumulative distribution function corresponds to a grayscale level of the original histogram. The function is applied to the original image data, and the grayscale value of each pixel is adjusted to achieve contrast enhancement. The new grayscale value of each pixel is obtained by modifying its original value by finding the CDF after equalization. The result generates an image preprocessing record, which includes the processed image, contrast and brightness values.

[0071] S102: Detecting and extracting edge information in the pathological slice image based on the image preprocessing record, and generating an edge information extraction record;

[0072] The specific formula for detecting and extracting edge information in pathological slice images is:

[0073]

[0074] Among them, G represents the edge strength value, G x Indicates the horizontal gradient of each pixel in the image, G y It represents the vertical gradient size of each pixel in the image. x represents the horizontal coordinate in the image, which is used to calibrate the position of the pixel in the image matrix. y represents the vertical coordinate in the image, which is used to calibrate the position of the pixel in the image matrix. c is the adjustment coefficient, which is used to adjust the weight ratio of gradient synthesis to the average value.

[0075] formula:

[0076]

[0077] Detailed explanation of the formula and the process of formula calculation and derivation:

[0078] The formula is used to calculate the edge strength of each pixel in the image, and the result is used to extract the edge information of the image;

[0079] Parameter meaning and setting value:

[0080] G x is the gradient size of the image in the horizontal direction, assuming that the horizontal gradient value of the target pixel is 6;

[0081] G yis the vertical gradient of the image, assuming that the vertical gradient value of the target pixel is 8;

[0082] c is the adjustment coefficient, assuming c is 0.5;

[0083] Substitute the parameters into the formula for calculation:

[0084] Compute the sum of squares part:

[0085]

[0086] Calculate the average gradient part:

[0087]

[0088] Calculate edge strength values:

[0089]

[0090] The result 13.5 indicates that the edge strength value of the current pixel is 13.5. The value reflects the composite gradient strength of the target point in the horizontal and vertical directions. The result is used to construct the edge strength matrix of the entire image and extract edge information.

[0091] S103: extracting records based on edge information, extracting color information and evaluating color distribution characteristics in the image to generate image edge and color information;

[0092] In sub-step S103, based on the edge information extraction record, the image is subjected to color information analysis and color distribution feature evaluation. The process uses a color histogram to identify and analyze the image color, reads the color value of each pixel in the image, and counts the frequency of occurrence of each color value. The range and density of the color distribution are observed through the color histogram, and the color distribution evaluation is performed, including calculating color statistics such as the mean and standard deviation, analyzing the main color components and their spatial distribution in the image. This step provides a data basis for subsequent image analysis such as pathological feature recognition. The generated image edge and color information includes a color distribution diagram and analysis results, and the specific statistical value of each color component is also recorded.

[0093] See also Figure 3 , based on the image edge and color information, the texture features of multiple locations in the image are extracted, including the texture directionality, roughness and contrast. Combined with the edge information, the shape features in the image are analyzed and extracted, the tissue structure is identified and the type information is recorded. The specific steps for generating tissue structure feature data are as follows:

[0094] S201: using image edge and color information, analyzing texture information of the pathological slice image, including measuring the directionality and roughness of textures at multiple locations, and generating texture analysis results;

[0095] In sub-step S201, texture analysis is performed to explore the texture features in the pathological slice image. The grayscale co-occurrence matrix method is used to measure the texture directionality and roughness of each area in the image. By calculating the frequency of occurrence of grayscale value pairs between pixels, a matrix is ​​formed to extract multiple texture descriptors, such as contrast, uniformity, homogeneity, and energy. The target descriptor helps to describe the coarseness and fineness of the image texture and obtain a texture characteristic description. The target data is aggregated into a texture analysis result, and the specific values ​​of the texture directionality and roughness of each area are recorded.

[0096] S202: Based on the texture analysis result and combined with the edge information, the shape features in the image are analyzed and extracted, the tissue structure is identified and marked, and a structure recognition map is generated;

[0097] In sub-step S202, a morphological-based method is used to identify and mark tissue structures, enhance the structural boundaries of tissues in the image, make the edges clearer, extract the precise boundaries of tissue structures through morphological gradients, remove small noise points, and apply contour tracking algorithms to determine and mark the contours of each independent tissue. The target operation results in a structure recognition map in which each tissue structure is clearly marked, including geometric position and shape description, providing accurate basic data for subsequent analysis.

[0098] S203: Analyze the shape of the tissue structure based on the structure recognition graph, identify the type of the tissue structure, and generate tissue structure feature data by combining the shape features and texture features of the tissue;

[0099] In sub-step S203, a support vector machine classifier is used to identify tissue types. SVM is used to classify data. Different types of tissue structures are separated by defining decision boundaries. During the classification process, shape and texture features are converted into vectors and input into the SVM model. The model learns and classifies according to the target features and outputs the type of each tissue. The target type information is integrated with the shape and texture data to generate tissue structure feature data. The target data records the classification results of each tissue structure and the related shape and texture descriptions, providing important quantitative information for medical diagnosis and research.

[0100] See also Figure 4 , using tissue structure feature data, by comparing the differences in texture and color at multiple locations in the tissue structure area, the lesion tissue is identified and the image is segmented. The specific steps for generating the segmented image of the lesion location are as follows:

[0101] S301: Analyze the color and texture parameters of the tissue area based on the tissue structure feature data, identify the lesion area by comparing and analyzing the color and texture information of multiple positions, and generate lesion calibration data;

[0102] In sub-step S301, an in-depth analysis of color and texture parameters is performed. The analysis uses the k-means clustering algorithm to process and compare the color and texture information of multiple locations in the image. K-means clustering is used to divide the data into multiple clusters and discover patterns in the data. By setting an appropriate number of clusters k, the algorithm iteratively calculates the center point of each cluster and the distance from each point to the center point, and divides each area. In the process, attention is paid to the similarities and differences in color and texture parameters to effectively distinguish normal and diseased tissues. The algorithm output includes the statistical characteristics of each cluster, such as the center color value and the average roughness of the texture. By comparing the target characteristics, the location of the diseased area is determined, and the lesion calibration data is generated. The data records the identified lesion area and characteristic parameters.

[0103] S302: Segmenting the lesion region based on the lesion calibration data, separating the normal tissue and the lesion tissue region, and generating lesion region segmentation data;

[0104] In sub-step S302, based on the obtained lesion calibration data, image segmentation technology, especially a threshold-based segmentation method, is used to separate the lesion tissue from the normal tissue. The threshold is determined based on the analysis of the lesion calibration data. By evaluating the differences in color and texture between the lesion tissue and the normal tissue, the optimal threshold is automatically set. After the threshold is applied, each pixel in the image is marked as lesion or normal, and the lesion area is segmented from the healthy tissue. The generated lesion area segmentation data includes the segmented image and information of each segmented area, such as the size, location and boundary description of the area. The target data is crucial for subsequent clinical diagnosis and treatment plan formulation.

[0105] S303: Based on the lesion area segmentation data, adjusting the segmentation boundary, optimizing the visibility of the lesion edge and the integrity of the lesion area, and generating a lesion position segmentation image;

[0106] In sub-step S303, the lesion area segmentation data is optimized, and the segmentation boundary is adjusted using morphological processing technology morphological filtering, including morphological dilation and erosion operations. The visibility of the lesion edge and the integrity of the lesion area are improved through target operations to make the lesion edge clearer. Erosion is used to remove small island structures caused by segmentation errors. Edge tracking technology is used to re-check and correct the segmented edges to ensure the accuracy and integrity of the lesion area. A lesion position segmentation image is generated through the target step. The image shows the optimized lesion area, marks the exact position and boundary of the lesion, and provides visual data for medical analysis.

[0107] See also Figure 5 , segment the image based on the lesion location, analyze the lesion tissue structure, calculate the geometric properties, including area, perimeter and complexity, analyze the lesion type by comparing the data features with the data of multiple known lesion types, and generate the lesion type analysis results in the following steps:

[0108] S401: Segment the image based on the lesion position, measure the area and perimeter of the target lesion area, and generate geometric attribute information;

[0109] In sub-step S401, the geometric property analysis of the segmented image of the lesion location is performed. The process uses the pixel counting method in the image processing software to determine the area. The number of pixels in the lesion area in the image is counted. Each pixel corresponds to the area in the actual tissue. The total area is calculated. The perimeter is measured using a boundary tracking method. The total perimeter is determined by tracking the outer edge pixels of the lesion area and accumulating the length of each edge segment. The target geometric data provides a basis for subsequent shape complexity analysis. The generated geometric property information includes the specific area and perimeter values ​​of each lesion area. The target information is recorded and analyzed to evaluate the size and morphological characteristics of the lesion area, providing quantitative data for clinical evaluation and further research.

[0110] S402: Based on the geometric attribute information, the shape complexity index of the lesion area is calculated by measuring the irregularity of the boundary of the lesion area and the length of the broken line, and the complexity analysis result is generated;

[0111] In sub-step S402, the shape complexity index of the lesion area is calculated by analyzing the irregularity of the boundary of the lesion area. The process uses the Fractal Dimension algorithm to evaluate the complexity of the shape. The fractal dimension is a quantitative indicator that describes the roughness of the shape. The fractal dimension analysis of the boundary of the lesion area is performed to measure the broken line length and irregularity of the boundary. The boundary is measured at different scales, and the coverage length of the boundary at each scale is calculated. The shape complexity index is obtained by linear regression analysis of the logarithm of the target length and the scale. The generated complexity analysis result records the shape complexity index of each lesion area. The index is used to understand the morphological characteristics and pathological type of the lesion.

[0112] S403: Based on the complexity analysis result, by comparing with geometric features of multiple known lesion types, identifying the lesion type in the target slice image, and generating a lesion type analysis result;

[0113] The specific formula for identifying the lesion type in the target slice image is:

[0114]

[0115] Among them, d represents the distance from the support vector to the decision boundary, n represents the total number of features involved in the calculation, i represents the index of the current feature, and w i represents the weight of the i-th feature, x i represents the value of the i-th feature, and b represents the bias term of the decision boundary.

[0116] formula:

[0117]

[0118] Detailed explanation of the formula and the process of formula calculation and derivation:

[0119] The formula is used to calculate the distance from the support vector to the decision boundary, and the result is used to identify the type of lesion in the image;

[0120] Parameter meaning and setting value:

[0121] n is the total number of features, assumed to be 4, reflecting the number of image features considered in the model, including area, perimeter, complexity, and shape;

[0122] w i is the weight of the ith feature, which is the average importance index obtained based on the analysis of historical lesion image data. Assume that w1 is the weight of the area feature, which is assumed to be 0.5, reflecting the high importance of area in lesion identification, because the area size is usually directly related to the severity of the lesion. w2 is the weight of the perimeter feature, which is assumed to be 0.3, reflecting the importance of the perimeter. The irregularity of the perimeter is usually related to certain types of lesions. w3 is the weight of the complexity feature, which is assumed to be 0.1, reflecting the auxiliary role of complexity in lesion identification. Complexity or morphological irregularity can assist in identifying specific types of lesions. w4 is the weight of the shape feature, which is assumed to be 0.1, reflecting the auxiliary diagnostic value of the shape feature. Specific changes in shape are also key indicators of lesions in some cases.

[0123] x i is the value of the ith feature, the measured data extracted from the target slice image by image processing software, x1 is the measured value of the area feature, assuming it is 120 square microns, reflecting the size of the lesion area directly measured from the image, x2 is the measured value of the perimeter feature, assuming it is 30 microns, reflecting the total length of the lesion boundary directly measured from the image, x3 is the measured value of the complexity feature, assuming it is 5, reflecting the irregularity of the lesion boundary calculated from the image, x4 is the measured value of the shape feature, assuming it is 3, reflecting the geometric parameters of the lesion morphology evaluated from the image

[0124] b is the bias term for classification decision, which is assumed to be -1.2, reflecting the classification boundary adjustment determined by the model during the optimization process on the training data set, which is used to increase the classification accuracy of the model for new samples;

[0125] Substitute the parameters into the formula for calculation:

[0126]

[0127] The calculated result d≈114.33 indicates the distance from the support vector to the classification decision boundary. The distance value is used to determine the degree of match between the lesion features in the target slice image and the known lesion type, helping to identify the lesion type.

[0128] See also Figure 6 According to the lesion type analysis results and the type information of the tissue structure, a three-dimensional model of the target tissue is established and the pathological slice image is mapped into the model, and various information is annotated. The specific steps for generating a three-dimensional pathological tissue model are as follows:

[0129] S501: Based on the lesion type analysis result and the type information of the target tissue structure, a three-dimensional model of the target tissue is constructed to obtain a model construction record;

[0130] In sub-step S501, computer-aided design software is used for three-dimensional modeling. The two-dimensional pathological slice image data is converted into a three-dimensional model using a voxelization method. During the process, the lesion type and tissue structure information in the image are used to automatically calculate the depth and structural hierarchy of the model, adjust the spatial scale and positioning of the model, and ensure the accuracy and detail reproduction of the model. The generated model construction record includes the modeling parameters and techniques used, the dimensions and size of the model, and a preview of the three-dimensional structure.

[0131] S502: Based on the model construction record, the pathological slice image is mapped into the model, and texture and color alignment is performed to obtain an image information mapping record;

[0132] In sub-step S502, based on the model construction record, the mapping of the pathological slice image to the three-dimensional model is performed. The texture mapping technology is used to ensure that the texture and color information of the image are accurately aligned to the model surface. The two-dimensional image is unfolded on the three-dimensional model surface using UV mapping. By adjusting the mapping parameters and correcting the color deviation, every detail in the image is copied to the model. The color fidelity and texture alignment are ensured in the process. The generated image information mapping record describes the technology used in the mapping process, the mapping parameters, and the visual effect of the model after mapping. The target record ensures the practicality and educational value of the model.

[0133] S503: Based on the image information mapping record, add annotation information to the model, including edge information, lesion area, and lesion type, to obtain a three-dimensional pathological tissue model;

[0134] In sub-step S503, annotation information is added to the three-dimensional model, and key medical information such as edge information, lesion area and lesion type is marked on the model using annotation tools. Medical three-dimensional visualization software such as VTK is used to add text descriptions to the corresponding parts of the model through the interface provided by the software, and marking lines and highlight areas are drawn to enhance the visual representation of the lesion area. Target annotation helps doctors understand the lesion structure. The generated three-dimensional pathological tissue model records include the location, content, and visual effects of the annotations, providing a learning tool for subsequent diagnosis and research.

[0135] See also Figure 7 , a pathological slice image analysis system, the pathological slice image analysis system is used to perform the above-mentioned pathological slice image analysis method, the system comprises:

[0136] The slice image processing module adjusts the brightness and contrast of the image based on the pathological slice image information, combines noise removal, and generates image preprocessing results;

[0137] The edge detection and color analysis module extracts edge information, analyzes the color distribution of the image, and generates image edge and color information based on the image preprocessing results;

[0138] The texture and shape analysis module analyzes the texture features in the image, including directionality and roughness, based on the image edge and color information, and extracts shape features. It combines the edge detection results to identify the tissue structure and generate tissue structure feature data;

[0139] The lesion region recognition module uses tissue structure feature data to identify the lesion region by comparing the texture and color information of multiple locations, and performs image segmentation on the lesion region to generate a lesion location segmentation image;

[0140] The lesion image analysis module segments the image based on the lesion location and calculates the geometric properties of the lesion area, including area, perimeter, and complexity. It analyzes the lesion type by comparing it with data of multiple known lesion types and generates lesion type analysis results.

[0141] The 3D modeling and mapping module constructs a 3D model of the target tissue based on the lesion type analysis results and the tissue type information in the slice image, maps the pathological slice image data to the model, performs information annotation, and generates a 3D pathological tissue model.

[0142] The above embodiments can be implemented in whole or in part by software, hardware (such as circuits), firmware or any other combination. When implemented by software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the process or function described in the embodiment of the present invention is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website, computer, server or data center to another website, computer, server or data center by wired (such as infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that contains one or more available media sets. The available medium can be a magnetic medium (for example, a floppy disk, a hard disk, a tape), an optical medium (for example, a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state hard disk.

[0143] It should be understood that the term "and / or" in this article is only a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. A and B can be singular or plural. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship, but it may also indicate an "and / or" relationship. Please refer to the context for specific understanding.

[0144] In the present invention, "at least one" means one or more, and "more than one" means two or more. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, at least one of a, b, or c can be represented by: a, b, c, ab, ac, bc, or abc, where a, b, c can be single or multiple.

[0145] It should be understood that in various embodiments of the present invention, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0146] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0147] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described equipment, devices and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0148] In the several embodiments provided by the present invention, it should be understood that the disclosed devices, apparatuses and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another device, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0149] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0150] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0151] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0152] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A pathological section image analysis method, characterized in that: The method comprises: Based on the pathological slice image information and combined with image preprocessing, the edge information in the image is detected and extracted, the color information is analyzed, the color distribution characteristics in the image are evaluated, and the image edge and color information is generated; Based on the image edge and color information, extract the texture features of multiple positions in the image, including the directionality, roughness and contrast of the texture, combine the edge information, analyze and extract the shape features in the image, identify the tissue structure and record the type information, and generate tissue structure feature data; Using the tissue structure feature data, by comparing the differences in texture and color at multiple locations in the tissue structure region, the diseased tissue is identified, and the image is segmented to generate a segmented image of the diseased location; Segment the image based on the lesion location, analyze the lesion tissue structure, calculate geometric properties, including area, perimeter, and complexity, analyze the lesion type by comparing the data features with data of multiple known lesion types, and generate a lesion type analysis result; According to the lesion type analysis results and the type information of the tissue structure, a three-dimensional model of the target tissue is established and the pathological slice image is mapped into the model, and various information is annotated to generate a three-dimensional pathological tissue model.

2. The pathological section image analysis method according to claim 1, characterized in that: The image edge and color information includes edge information data set, color extraction information, and image preprocessing parameters; the tissue structure feature data includes texture information data set, shape analysis results, and tissue structure type labels; the lesion position segmentation image includes lesion area boundaries, texture feature mapping data within the lesion area, and a lesion area color distribution map; the lesion type analysis results include lesion type recognition results, lesion area geometric properties, and matching degree calculation results; the three-dimensional pathological tissue model includes three-dimensional model construction results, annotation information, and slice image mapping results.

3. The pathological section image analysis method according to claim 1, characterized in that: Based on the pathological slice image information, combined with image preprocessing, the edge information in the image is detected and extracted, the color information is analyzed, and the color distribution characteristics in the image are evaluated. The specific steps for generating image edge and color information are as follows: Based on the pathological slice image information, the image is preprocessed, the visibility of the edge information in the image is optimized by adjusting the contrast and brightness of the image, and an image preprocessing record is generated; Based on the image preprocessing record, detecting and extracting edge information in the pathological slice image, and generating an edge information extraction record; Based on the edge information extraction record, color information is extracted and color distribution characteristics in the image are evaluated to generate image edge and color information.

4. The pathological section image analysis method according to claim 3, characterized in that: The specific formula for detecting and extracting edge information in the pathological slice image is: Among them, G represents the edge strength value, G x Indicates the horizontal gradient of each pixel in the image, G y It represents the vertical gradient size of each pixel in the image. x represents the horizontal coordinate in the image, which is used to calibrate the position of the pixel in the image matrix. y represents the vertical coordinate in the image, which is used to calibrate the position of the pixel in the image matrix. c is the adjustment coefficient, which is used to adjust the weight ratio of gradient synthesis to the average value.

5. The pathological section image analysis method according to claim 1, characterized in that: Based on the image edge and color information, texture features at multiple locations in the image are extracted, including texture directionality, roughness and contrast. In combination with edge information, shape features in the image are analyzed and extracted, tissue structure is identified and type information is recorded. The steps of generating tissue structure feature data are specifically as follows: Using the image edge and color information, analyzing the texture information of the pathological slice image, including measuring the directionality and roughness of the texture at multiple locations, and generating a texture analysis result; Based on the texture analysis results, combined with edge information, shape features in the image are analyzed and extracted, tissue structures are identified and marked, and a structure recognition map is generated; Based on the structure recognition map, the shape of the tissue structure is analyzed to identify the type of the tissue structure, and the tissue structure feature data is generated by combining the shape features and texture features of the tissue.

6. The pathological section image analysis method according to claim 1, characterized in that: Using the tissue structure feature data, by comparing the differences in texture and color at multiple locations in the tissue structure area, the diseased tissue is identified, and the image is segmented. The steps of generating a segmented image of the lesion location are specifically as follows: Based on the tissue structure feature data, analyzing the color and texture parameters of the tissue area, identifying the lesion area by comparing and analyzing the color and texture information of multiple positions, and generating lesion calibration data; Based on the lesion calibration data, the lesion area is segmented to separate the normal tissue and the lesion tissue area, and the lesion area segmentation data is generated; Based on the lesion region segmentation data, the segmentation boundary is adjusted to optimize the visibility of the lesion edge and the integrity of the lesion region, and a lesion position segmentation image is generated.

7. The pathological section image analysis method according to claim 1, characterized in that: The steps of segmenting the image based on the lesion position, analyzing the lesion tissue structure, calculating geometric properties including area, perimeter and complexity, analyzing the lesion type by comparing the data features with data of multiple known lesion types, and generating the lesion type analysis result are as follows: Based on the lesion position segmentation image, the area and perimeter of the target lesion region are measured to generate geometric attribute information; Based on the geometric attribute information, by measuring the irregularity of the boundary of the lesion area and the length of the broken line, the shape complexity index of the lesion area is calculated to generate a complexity analysis result; Based on the complexity analysis result, the lesion type in the target slice image is identified by comparing with the geometric features of multiple known lesion types, and a lesion type analysis result is generated.

8. The pathological section image analysis method according to claim 7, characterized in that: The specific formula for identifying the lesion type in the target slice image is: Among them, d represents the distance from the support vector to the decision boundary, n represents the total number of features involved in the calculation, i represents the index of the current feature, and w i represents the weight of the i-th feature, x i represents the value of the i-th feature, and b represents the bias term of the decision boundary.

9. The pathological section image analysis method according to claim 1, characterized in that: According to the lesion type analysis results and the type information of the tissue structure, a three-dimensional model of the target tissue is established and the pathological slice image is mapped into the model, and various information is annotated. The specific steps of generating a three-dimensional pathological tissue model are as follows: Based on the lesion type analysis result and according to the type information of the target tissue structure, a three-dimensional model of the target tissue is constructed to obtain a model construction record; Based on the model construction record, the pathological slice image is mapped into the model, and the texture and color are aligned to obtain an image information mapping record; Based on the image information mapping record, annotation information is added to the model, including edge information, lesion area, and lesion type, to obtain a three-dimensional pathological tissue model.

10. A pathological section image analysis system, characterized in that: According to the pathological section image analysis method according to any one of claims 1 to 9, the system comprises: The slice image processing module adjusts the brightness and contrast of the image based on the pathological slice image information, combines noise removal, and generates image preprocessing results; The edge detection and color analysis module extracts edge information, analyzes the color distribution of the image, and generates image edge and color information based on the image preprocessing result; The texture and shape analysis module analyzes the texture features in the image, including directionality and roughness, based on the image edge and color information, and extracts shape features, combines the edge detection results, identifies the tissue structure, and generates tissue structure feature data; The lesion region recognition module uses the tissue structure feature data to identify the lesion region by comparing the texture and color information of multiple positions, and performs image segmentation on the lesion region to generate a lesion position segmentation image; The lesion image analysis module segments the image based on the lesion location and calculates the geometric properties of the lesion area, including area, perimeter, and complexity. It analyzes the lesion type by comparing it with data of multiple known lesion types and generates lesion type analysis results. The three-dimensional modeling and mapping module constructs a three-dimensional model of the target tissue according to the lesion type analysis results and the type information of the tissue in the slice image, maps the pathological slice image data to the model, performs information annotation, and generates a three-dimensional pathological tissue model.

Images