Precise recognition method and system for typical area of pathological image

Abstract

The invention provides a pathological image typical area accurate identification method and system. The pathological image typical area accurate identification method comprises the following steps: acquiring a pathological image, and screening out a preliminary typical area of the pathological image; and inputting the obtained preliminary typical region into a pathological image accurate analysis network, carrying out image feature separation on the preliminary typical region image, and re-fusing the image features to realize accurate recognition of the typical region. According to the method, the feature separation module and the feature re-fusion module are combined, so that the output typical region is more accurate, and the recognition precision of the fine granularity of the image is higher and more accurate.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence, and in particular to a method and system for accurately identifying typical areas of pathological images. Background Art

[0002] Fine-grained image recognition is to identify different subcategories from the same large category. Most deep learning-based fine-grained image recognition methods extract regional features by reusing the backbone network, and improve the network's ability to capture subtle differences by locating the most discriminative local areas. However, since the key areas are too subtle, it is impossible to obtain all the key information through traditional convolutional networks. Researchers have begun to work on improvements within the convolutional network framework to further locate key areas and enrich feature representations. At present, fine-grained image recognition methods are mainly divided into two categories. The first type of method explicitly locates fine-grained key areas and then fuses the located areas. The second type is an implicit method that uses high-order features to represent end-to-end feature encoding or specially designed loss functions to drive the model to discover fine-grained key areas that reflect subtle differences.

[0003] However, there is both positive and negative feature transfer between image data of different categories. Specifically, when using existing methods for fine-grained image recognition, after joint training of image data of different categories, although the overall distribution of features of each category of image data is almost always better than that of separate training (i.e., larger inter-class variance and intra-class similarity), it is obvious that only some categories of images get better feature representation, while other categories of images get worse feature representation. At the same time, the feature boundaries between image data of different categories sometimes become more blurred, confusing the recognition performance of the model between them. However, negative feature transfer is dominant in practice, resulting in a significant decrease in the test accuracy of the model on a single category image dataset compared to separate training. Summary of the invention

[0004] In order to overcome the defects in the above-mentioned prior art, the object of the present invention is to provide a method and system for accurately identifying typical areas of pathological images.

[0005] In order to achieve the above-mentioned object of the present invention, the present invention provides a method for accurately identifying typical areas of pathological images, comprising the following steps:

[0006] Acquire pathological images;

[0007] Screening of preliminary typical areas of pathological images;

[0008] The obtained preliminary typical regions are input into the precise analysis network of pathological images, and image features of the preliminary typical region images are separated and the image features are reintegrated to achieve precise identification of typical regions.

[0009] This method combines the feature separation module and the feature re-fusion module to make the output typical area more precise and the fine-grained recognition accuracy of the image higher and more accurate.

[0010] In an optional scheme of the method for accurately identifying typical areas of pathological images, the screening method of the preliminary typical areas is one of the uniform blocking method, the Overlap blocking method, the analysis method of multi-instance deep learning, the analysis method based on self-supervised learning, and the hierarchical sampling network CMLS analysis method.

[0011] In an optional solution of the method for accurately identifying typical areas of pathological images, the pathological image accurate analysis network includes a feature separation module and a feature re-fusion module;

[0012] The feature separation module separates the image features of the preliminary typical region image by a feature extraction unit specific to the image category;

[0013] The feature re-fusion module re-fuses the image features based on a channel selection mechanism to achieve accurate recognition of typical areas in the preliminary typical area image.

[0014] In this optional solution, the feature separation module can alleviate the negative feature transfer generated by images of different categories, and the feature re-fusion module can enhance the positive feature transfer of specific image categories.

[0015] In an optional solution of the method for accurately identifying typical regions of pathological images, the feature separation module uses multiple category feature maps to obtain specific category features, and uses specific category classifiers to constrain different category features to be separable.

[0016] In an optional solution of the method for accurately identifying typical areas of pathological images, the expression of the accurate analysis network of pathological images is AAOH(·)=AAOH-FSM(·)⊕AAOH-FRM(·),

[0017] AAOH-FSM(·) is the action expression of the feature separation module.

[0018]

[0019] Among them, f n Represents the features of the nth image category, image features after separation from common features x n is a sample of the nth image category, Conv n (·) A 1×1 convolution is used to extract the specific features of the nth image category, with the input and output channels being the same as f n The number of feature channels is the same, BN n (·) is a batch normalization layer, ReLU is a nonlinear activation function, Gn (·) is the classifier for the nth image category, is a set of features of a specific image category, and P is the total number of image categories;

[0020] AAOH-FRM(·) is the action expression of the feature re-fusion module. Represents the reintegration characteristics, express The τth th channels, τ∈[1,T], the total number of channels is T, Cat(·) represents the concatenation of multiple features at the channel level, FC(·) represents the fully connected layer with an output dimension equal to the image category n,

[0021] The fused features F are passed to the joint classifier G(·) to predict the joint classification result.

[0022] In this optional solution, the feature separation module decouples mixed features, embeds specific features of different categories, and eliminates feature negative transfer; the feature re-fusion module fuses feature arrays based on a channel selection mechanism to enhance feature positive transfer of specific image categories.

[0023] In an optional solution of the method for accurately identifying typical areas of pathological images, the loss function of the pathological image accurate analysis network is Loss all (·)=Loss FSM (·)+Loss FRM (·),

[0024] Loss FSM (·) is the loss function of the feature separation module:

[0025]

[0026] Loss FRM (·) is the loss function of the feature re-fusion module: Loss FRM (·)=L(G(F),y), where L(·) is the cross entropy loss, y represents the true category, and G(F) represents the category predicted by the module.

[0027] In this optional solution, the precise analysis network of ovarian tissue pathology images accurately identifies the input typical image area, and it is optional in the workflow of the entire prognostic analysis system, wherein: the feature separation module separates the image features of the preliminary typical area image through a feature extraction unit specific to the image category, thereby alleviating the negative migration of features generated by images of different categories; the feature re-fusion module re-fuses the image features specific to the image category based on the channel selection mechanism, thereby enhancing the positive migration of features of the specific image category.

[0028] The present invention also proposes a pathological image recognition system, including an image receiving module, a processing module and a storage module, wherein the image receiving module receives a pathological image and sends the received pathological image to the processing module, wherein the processing module is in communication connection with the storage module, and the storage module is used to store at least one executable instruction, wherein the executable instruction enables the processing module to perform operations corresponding to the above-mentioned pathological image typical area accurate recognition method according to the image received by the processing module. The system has all the advantages of the above-mentioned method.

[0029] Additional aspects and advantages of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] The above and / or additional aspects and advantages of the present invention will become apparent and easily understood from the description of the embodiments in conjunction with the following drawings, in which:

[0031] Figure 1 Schematic diagram of the structure of a configurable hierarchical sampling network;

[0032] Figure 2 It is a schematic diagram of the structure of the j-th level sampling module in the configurable hierarchical sampling network;

[0033] Figure 3 It is a structural diagram of the precise analysis network for pathological images. In the figure, an image containing three image categories is taken as an example. Sub-tiles is a preliminary typical area image. F(.) represents the feature set of the three image category features (f1(), f2(), f3()). Feature Separation is a feature separation module. G1(·), G2(·), and G3(·) are classifiers for the three image categories respectively. Feature Refusion is a feature refusion module. G(·) is a joint classifier. Fine-grained sub-tiles are highly fine-grained typical area images. DETAILED DESCRIPTION

[0034] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and cannot be understood as limiting the present invention.

[0035] In the description of the present invention, unless otherwise specified and limited, it should be noted that the terms "installed", "connected" and "connected" should be understood in a broad sense. For example, it can be a mechanical connection or an electrical connection, or it can be the internal connection between two components. It can be a direct connection or an indirect connection through an intermediate medium. For ordinary technicians in this field, the specific meanings of the above terms can be understood according to the specific circumstances.

[0036] Embodiment 1

[0037] The present invention provides an embodiment of a method for accurately identifying typical areas of pathological images, which specifically includes the following steps:

[0038] Obtain a pathological image and screen preliminary typical areas of the pathological image. For pathological images of regular size, screening preliminary typical areas of the pathological image can be performed by using one of the uniform block method, the overlap block method, the multi-instance deep learning analysis method, and the self-supervised learning-based analysis method.

[0039] This embodiment uses pathological images of ovarian HGSA (High-grade serous adenocarcinomas) patients as samples. Each pathological image is a full-field pathological image of the primary lesion of ovarian tumor tissue.

[0040] Since the resolution of the pathological full-field image is too large (reaching the billion or even billion pixels), in this embodiment, the pathological full-field image is input into a configurable hierarchical sampling network CMLS that has been pre-trained according to the screening requirements. The configurable hierarchical sampling network CMLS is used to perform hierarchical sampling on the pathological full-field image to screen out preliminary typical areas, and the obtained preliminary typical areas are input into the pathological image accurate analysis network AAOH that has been pre-trained according to the fine-grained requirements to achieve accurate recognition of curved areas of the preliminary typical area image.

[0041] The configurable layered sampling network CMLS includes cascaded multi-stage sampling modules. The structure of the configurable layered sampling network CMLS in this embodiment is as follows: Figure 1 As shown, it includes the first-level sampling module CMLS-SAP1 (·), the second-level sampling module CMLS-SAP2 (·), and the N-level sampling module CMLS-SAP n (·). The number N of sampling modules is configurable. In this embodiment, the value range of N is 1-5, preferably 2 or 3.

[0042] The first-level sampling module CMLS-SAP1(·) performs preliminary typical area screening on the low-resolution image of the pathological full-field image and maps it to the pathological full-field image of the original resolution, and obtains TN1 first-level sub-image blocks of the original resolution with the same area range as the screened area, where TN1 is an integer greater than 1;

[0043] The sampling modules at each level are cascaded in sequence and perform preliminary typical area screening on the low-resolution image of the sub-image block output by the sampling module at the previous level and map it to the pathological full-field image of the original resolution, thereby obtaining multiple sub-image blocks of each level with the same area range as the currently screened area.

[0044] That is, the second-level sampling module CMLS-SAP2(·) performs preliminary typical area screening on the low-resolution image of TN1 first-level sub-image blocks and maps the original resolution to the pathological full-field image, and obtains TN2 second-level sub-image blocks with the same original resolution as the area range of the current screened area, where TN2 is an integer greater than 1.

[0045] Sequentially, the j-th level sampling module CMLS-SAP j (·) Perform preliminary typical area screening on the low-resolution image of the sub-image block output by the previous sampling module and map it to the pathological full-field image of the original resolution to obtain multiple j-th level sub-image blocks of the original resolution with the same area range as the currently screened area, where j is an integer greater than 1.

[0046] Until the Nth level sampling module CMLS-SAP n (·) In TN n-1 The initial typical area screening is performed on the low-resolution image of the (N-1)th level sub-image block and mapped to the original resolution pathological full-field image to obtain the original resolution TN with the same area range as the currently screened area. N Nth level sub-image blocks, these sub-image blocks are the preliminary typical region images, which contain image detail information for subsequent analysis, TN N is an integer greater than 1.

[0047] The low resolutions collected by different sampling modules in the configurable layered sampling network CMLS can be the same or different, all between 0 and 1, preferably but not limited to increasing step by step, emphasizing the ability to process a small number of full-resolution image blocks of the input image, which can significantly reduce the peak GPU memory usage and have higher sampling accuracy. For very large images (such as mega- to gigapixel), the number of selected image blocks is much smaller than the sample space || S sampling ||'s size.

[0048] like Figure 2As shown, each sampling module of the configurable hierarchical sampling network CMLS includes a feature extraction unit CMLS-SAP j -FE(·), attention weighted unit CMLS-SAP j -AM(·) and taxon CMLS-SAP j -CF(·), (1≤j≤N). Feature extraction unit CMLS-SAP j -FE(·) extracts features from the full-field pathology image or the sub-image block output by the previous sampling module, and the classification unit CMLS-SAP j -CF(·) is based on the attention weighted unit CMLS-SAP j The weight of -AM(·) is used to classify the features extracted by the feature extraction unit to obtain the sub-image blocks.

[0049] In this embodiment, the feature extraction unit CMLS-SAP j -FE(·) is composed of trainable parameters A typical example of a deep convolutional neural network for characterization is shown in Table 1.

[0050] Attention Weighted Unit CMLS-SAP j -AM(·) is composed of trainable parameters A smaller deep convolutional neural network represented by is used to obtain the attention weight matrix, as a typical example of which is shown in Table 2.

[0051] Classification unit CMLS-SAP j -CF(·) is composed of trainable parameters A single fully connected layer fc-n representing class .

[0052] Table 1 Feature extraction unit CMLS-SAP j A typical example of -FE(·)

[0053] layer type 1 Conv(3,1,1,8)+ReLU() 2 Conv(3,1,1,16)+ReLU() 3 Conv(3,1,1,32)+ReLU() 4 Conv(3,1,1,16)+ReLU() 5 Conv(3,1,1,8)+ReLU() 6 Conv(3,1,1,16)+ReLU() 7 Conv(3,1,1,32)+ReLU() 8 Conv(3,1,1,1)+ReLU() 9 GlobalAveragePooling2D() 10 SoftMax()

[0054] Table 2 Attention weighted unit CMLS-SAP j A typical example of -AM(·)

[0055] layer type 1 Conv(3,1,1,8)+ReLU() 2 Conv(3,1,1,16)+ReLU() 3 Conv(3,1,1,32)+ReLU() 4 Conv(3,1,1,1)+SoftMax() 5 SoftMax()

[0056] The action function CMLS-SAP1(x) of the first level (i.e., when j=1) sampling module of the configurable hierarchical sampling network CMLS is as follows, which is given by θ={θ f ,θ a ,θ c}Parameterized convolutional neural network:

[0057]

[0058] in, is the classification function of the classification unit of the first-level sampling module, is the feature extraction function of the feature extraction unit of the first-level sampling module, and the sampling function From the input image x1∈R H×W Extract the first-level sub-image block of size h1×w1 from the image, which corresponds to the low-resolution view V(x,s1)∈R with a scale of s1∈(0,1) h×w The coordinates c1 = {i, j} in the sampling function, c1 represents the corresponding coordinates of the thumbnail view (the low-resolution view after the original image (pathological full-field image) is reduced) processed by the first-level sampling module and the original image (pathological full-field image), (i, j) is its horizontal and vertical coordinates in the original image, and the sampling function pass Maps coordinate c1 to a position in image x1 and returns a sub-image patch of size h1×w1, where H×W is the resolution of image x1. and h is the height of the thumbnail view of the input level 1 sampling module, w is the width of the reduced view of the input level 1 sampling module, It is a floor operation.

[0059] The attention mechanism is introduced in CMLS-SAP1(x), and the convolutional neural network of the attention mechanism is α=a θ (V(x,s1)), where a θ (·) represents the trainable parameters The smaller deep convolutional neural network represented is used to obtain the attention weight matrix and obtain the attention weighted network:

[0060] Among them, g Θ Generally refers to the classification function in the sampling module, here refers to the classification function in the l-th level sampling module f Θ Generally refers to the feature extraction function in the sampling module, here refers to the feature extraction function in the l-th layer sampling module is the attention matrix, c = (i, j) represents the corresponding coordinates of the thumbnail view processed by the sampling module of this layer and the original image (pathological full-field image), (i, j) represents its horizontal and vertical coordinates in the original image, C is the set of c in the sampling modules of each layer, α c Convolutional neural network with attention mechanism α=a θ (V(x,s1,c)),||S sampling|| is the set of all index pairs of the low-resolution view V(x,s1) of image x1, ||S sampling ||=h×w.

[0061] In order to avoid computing features from all sub-image patches of the low-resolution view V(x,s1), we can modify the formula Perform Monte Carlo approximation to obtain the function of the first-level sampling module: Where C = {(i,j)~a θ (V(x, s1))|i=1,2,...,M} is the index pair M<<||S of the low-resolution view V(x, s1) sampling ||. CMLS-SAP1(x) uses the low-resolution view V(x,s1) of image x1 to calculate the attention weights, obtains M sub-image blocks by sampling, and The average is used to obtain the aggregated features, P refers to the total number of image categories, z n Refers to the features unique to the nth image category.

[0062] Similarly, the action function of the j-th level sampling module of the configurable hierarchical sampling network CMLS is is given by θ={θ f ,θ a ,θ c}Parameterized convolutional neural network.

[0063] in, is the classification function of the classification unit of the j-th level sampling module, is the feature extraction function of the feature extraction unit of the j-th level sampling module, View V(x,s j ,c)∈R u×v is the sampling function Based on scale S j ∈(0,1), c can map the thumbnail view of the layer back to the original image (pathological full-field image). CMLS-SAP j (x) introduces the view V(x,s j ,c) Attention mechanism CMLS-SAP j -AM(·), CMLS-SAP j -AM(·)=V(x,s j ,c)∈R u×v , we get the attention weighted network:

[0064] Where x is the input image, is the attention function CMLS-SAP of the j-th level sampling module j -AM(·), the j-1th level sampling function is the input image x to the j-1th level sampling module j-1 The ratio is S j-1 ∈(0,1) low-resolution sampling to obtain the coordinate c (j-1) And c (j-1) Mapped to the j-1th level sampling module input image x j-1 middle, is the j-th level sampling function, the input image x of the j-th level sampling module j for c (j) is the input image x of the j-th level sampling module j The ratio is S j The low-resolution sampling of ∈(0,1) corresponds to the coordinate mapping of the pathological full-field image, and C is the coordinate mapping of all c (j) The set composed of.

[0065] The low-resolution sampling ratio of the input image for each level of sampling function is set in advance.

[0066] right Perform Monte Carlo approximation to obtain the function of the j-th level sampling module:

[0067]

[0068] Where M refers to the total number of sub-image blocks output by each sampling module, c (j-1) ∈C, c (j) ~CMLS-SAP j-1 (V(x,s j-1 ,c (j-1) ))From the distribution V(x,s j-1 ,c (j-1) ) corresponding position extraction, CMLS-SAP j-1 (·) represents the function of the j-1th level sampling module, CMLS-SAP j (x) Use image x j A low-resolution or full-resolution sub-image block from and c (j) The j-level discrete distribution is determined by the j-level low-resolution view V(x,s j ,c (j) ) is obtained. When j = 1, CMLS-SAP j (x) is the selection of typical feature areas on the low-resolution image; subsequently, the typical feature areas are magnified step by step and then sampled until j=N, and then the full-resolution image (pathological full-field image) is performed.

[0069] The loss function of the configurable layered sampling network CMLS is:

[0070]

[0071] in, is the weighted sum of the cross entropy of each level of sampling modules, L RZ is the regularization term, β is the weight coefficient of the regularization term, and the default value is 1.0.

[0072] The weighted sum of the cross entropy of each level of sampling modules Contains the cross entropy of the j-th level sampling module and the weight α of the j-th level sampling module (j) . Weight α (j) It consists of two parts, including the area ratio ω1 of the output / input image of the j-th level sampling module (j) , and the difference between the ratio of the sum of the gradient values ​​of the image block with the smallest sum of gradient information and the image block with the largest sum of gradient information among all the image blocks at the jth level of the output and 1

[0073] Among them, M (j) The total number of image blocks output by the j-th level sampling module,

[0074] Among them, tile (j) is the gradient information in the image block output by the j-th sampling module, and m is the sequence number of the image block output by the j-th sampling module.

[0075] Weight ω1 (j) Represents the area ratio of the output / input image of the j-th sampling module; the smaller the area ratio, the larger the weight is; the ratio of the number of image blocks output by the j-th sampling module to the sum of the number of image blocks output by all sampling modules can be directly used to achieve fast approximate calculation. Weight ω2 (j) Represents the ratio of the minimum value to the maximum value of the gradient information in the output image block of the current j-th layer sampling module. The larger the value, the higher the degree of information consistency of the output sampling image block, and a larger weight is given to this part of the output.

[0076] The purpose of introducing a regularization term into the loss function is to keep the model simple while minimizing the training error, while improving the generalization performance of the model and preventing overfitting.

[0077] L RZ =max{||CMLS-SAP1-AM(·)||2,||CMLS-SAP2-AM(·)||2,...,||CMLS-SAP N -AM(·)||2}, where ||CMLS-SAP j−AM(·)||2 is the second-order norm of the network weights of the attention unit of the jth sampling module.

[0078] The structure of the pathological image accurate analysis network AAOH is as follows Figure 3 As shown, the feature separation module AAOH-FSM and the feature fusion module AAOH-FRM.

[0079] The feature separation module AAOH-FSM separates the image features of the preliminary typical area image through a feature extraction unit specific to the image category, thereby alleviating the negative feature transfer produced by images of different categories; the feature re-fusion module AAOH-FRM re-fuses the image features based on the channel selection mechanism, realizes accurate recognition of typical areas in the preliminary typical area image, and enhances the positive feature transfer of specific image categories.

[0080] The function expression of the pathological image accurate analysis network AAOH is AAOH(·)=AAOH-FSM(·)⊕AAOH-FRM(·), and the loss function of the pathological image accurate analysis network is Loss all (·)=Loss FSM (·)+Loss FRM (·),

[0081] Among them, AAOH-FSM(·) is the action expression of the feature separation module AAOH-FSM, Loss FSM (·) is the loss function of the feature separation module AAOH-FSM, AAOH-FRM(·) is the function expression of the feature fusion module AAOH-FRM, Loss FRM (·) is the loss function of the feature fusion module AAOH-FRM.

[0082] The feature separation module AAOH-FSM decouples mixed features, embeds specific features of different categories, eliminates feature negative transfer, uses multiple category feature maps to obtain specific category features, and uses specific category classifiers to constrain different category features to be separable.

[0083] The action expression of the feature separation module AAOH-FSM is:

[0084] Among them, f n The features of the nth image category. The category of the specific image can be obtained through the training of the feature separation module, and the image features after separation of the common features x n is a sample of the nth image category, Conv n (·) A 1×1 convolution is used to extract the specific features of the nth image category, with the input and output channels being the same as f n The number of feature channels is the same, BNn (·) is a batch normalization layer, ReLU is a nonlinear activation function, G n (·) is the classifier for the nth image category, which is only used in the training phase and will be discarded in the inference phase. is a set of features of a specific image category, and P is the total number of image categories.

[0085] Loss function of feature separation module AAOH-FSM FSM (·)for:

[0086] L refers to the cross entropy operation, Refers to the classification head for features The predicted output value of n Refers to the nth image block of a specific category, which is GroundTruth; Pointer pair y n Find the cross entropy.

[0087] The feature re-fusion module AAOH-FRM(·) fuses feature arrays based on the channel selection mechanism, and its function expression is:

[0088] in, Represents the reintegration characteristics, express The τth th channels, τ∈[1,T], the total number of channels is T, Cat(·) represents the cascade of multiple feature channels, and FC(·) represents the fully connected layer with an output dimension equal to the image category n. Finally, the fusion feature F τ Passed to the joint classifier G(·) to predict the joint classification result.

[0089] Loss function of feature fusion module AAOH-FRM FRM (·) is: Loss FRM (·)=L(G(F),y), where L(·) is the cross entropy loss, y represents the true category, and G(F) represents the category predicted by the module.

[0090] The present invention also provides a pathological image recognition system, including an image receiving module, a processing module and a storage module, wherein the image receiving module receives pathological images and sends the received pathological images to the processing module, the processing module is communicatively connected to the storage module, and the storage module is used to store at least one executable instruction, wherein the executable instruction enables the processing module to perform operations corresponding to the above-mentioned method for accurately identifying typical areas of pathological images according to the image it receives.

[0091] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "examples", "specific examples", or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.

[0092] Although the embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the claims and their equivalents.

Claims

1. A method for accurately identifying typical areas of pathological images, characterized in that: The following steps are involved: Acquire pathological images; Screening of preliminary typical areas of pathological images; The obtained preliminary typical regions are input into the precise analysis network of pathological images, and image features of the preliminary typical region images are separated and the image features are reintegrated to achieve precise identification of typical regions.

2. The method for accurately identifying typical areas of pathological images according to claim 1, characterized in that: The screening method for the preliminary typical areas is one of the uniform blocking method, the Overlap blocking method, the multi-instance deep learning analysis method, the self-supervised learning-based analysis method, and the hierarchical sampling network CMLS analysis method.

3. The method for accurately identifying typical areas of pathological images according to claim 1, characterized in that: The pathological image accurate analysis network includes a feature separation module and a feature re-fusion module; The feature separation module separates the image features of the preliminary typical region image by a feature extraction unit specific to the image category; The feature re-fusion module re-fuses the image features based on a channel selection mechanism to achieve accurate recognition of typical areas in the preliminary typical area image.

4. The method for accurately identifying typical areas of pathological images according to claim 3, characterized in that: The feature separation module uses multiple category feature maps to obtain specific category features, and uses specific category classifiers to constrain different category features to be separable.

5. The method for accurately identifying typical areas of pathological images according to claim 3, characterized in that: The expression of the pathological image accurate analysis network is: AAOH-FSM(·) is the action expression of the feature separation module. Among them, f n Represents the features of the nth image category, image features after separation from common features x n is a sample of the nth image category, Conv n (·) A 1×1 convolution is used to extract the specific features of the nth image category, with the input and output channels being the same as f n The number of feature channels is the same, BN n (·) is a batch normalization layer, ReLU is a nonlinear activation function, G n (·) is the classifier for the nth image category, is a set of features of a specific image category, and P is the total number of image categories; AAOH-FRM(·) is the action expression of the feature re-fusion module. Represents the reintegration characteristics, express The τth th channels, τ∈[1,T], the total number of channels is T, Cat(·) represents the concatenation of multiple features at the channel level, FC(·) represents the fully connected layer with an output dimension equal to the image category n, The fusion feature F τ Passed to the joint classifier G(·) to predict the joint classification result.

6. The tumor prognosis analysis method according to claim 3, characterized in that: The loss function of the pathological image accurate analysis network is Loss all (·)=Loss FSM (·)+Loss FRM (·), Loss FSM (·) is the loss function of the feature separation module: Loss FRM (·) is the loss function of the feature re-fusion module: Loss FRM (·)=L(G(F),y), where L(·) is the cross entropy loss, y represents the true category, and G(F) represents the category predicted by the module.

7. A pathological image recognition system, characterized in that: It includes an image receiving module, a processing module and a storage module, the image receiving module receives a pathological image and sends the received pathological image to the processing module, the processing module is communicatively connected to the storage module, the storage module is used to store at least one executable instruction, the executable instruction enables the processing module to perform an operation corresponding to the method for accurately identifying typical areas of pathological images as described in any one of claims 1-6 according to the image it receives.

Images