Pathological slice image processing method, electronic device and storage medium

Abstract

The present invention provides a pathological slice image processing method, electronic device and storage medium, the method comprising: obtaining a pathological slice image to be processed; classifying the pathological slice image to be processed to obtain an initial tumor classification result of the pathological slice image to be processed; identifying whether there is hyperplasia and necrosis in the pathological slice image to be processed to obtain a hyperplasia and necrosis recognition result of the pathological slice image to be processed; and obtaining a target tumor classification result of the pathological slice image to be processed according to the initial tumor classification result and the hyperplasia and necrosis recognition result. The present invention can solve the problems of time-consuming and labor-intensive and low accuracy caused by manual identification of brain glioma in the prior art.

Description

Technical Field

[0001] The present invention relates to the technical field of medical image processing, and in particular to a pathological slice image processing method, electronic equipment and storage medium. Background Art

[0002] Gliomas are the most common primary brain malignant tumors caused by the canceration of glial cells in the brain and spinal cord. Their incidence rate accounts for about 35.2% to 61.0% of intracranial tumors. They are derived from glial cells and have the characteristics of high incidence, high recurrence rate, high mortality rate and low cure rate. Gliomas are divided into multiple subtypes, and can be divided into astrocytomas and oligodendrocytomas according to the morphology of glioma cells.

[0003] At present, the classification of brain gliomas requires manual identification of pathological slice images. However, pathological slice images are not just one or a few images. In order to make a pathological diagnosis, a microscope is usually used to scan the entire pathological slice in a Z-shape under a 20x or 40x objective lens. A full pathological slice often scans thousands of slice images. Traditional manual identification methods require pathologists to browse through these thousands of images in detail before giving a diagnosis. This is not only time-consuming and labor-intensive, but also during the browsing process, some non-obvious high-risk features may be overlooked, thus affecting the accuracy of the diagnosis results. Summary of the invention

[0004] In order to solve the problems of time-consuming, labor-intensive and low accuracy in the manual identification of brain glioma in the prior art, the present invention provides a pathological slice image processing method, an electronic device and a storage medium.

[0005] In order to achieve the above object, the present invention adopts the following technical solution:

[0006] In a first aspect, a pathological slice image processing method is provided, comprising:

[0007] Acquire the pathological slice image to be processed;

[0008] Performing classification processing on the pathological slice image to be processed to obtain an initial tumor classification result of the pathological slice image to be processed;

[0009] Identify whether there is hyperplasia and necrosis in the pathological slice image to be processed, and obtain a hyperplasia and necrosis identification result of the pathological slice image to be processed;

[0010] According to the initial tumor classification result and the hyperplasia and necrosis identification result, a target tumor classification result of the pathological slice image to be processed is obtained.

[0011] In a preferred embodiment of the present invention, the classifying process of the pathological slice image to be processed to obtain the initial tumor classification result of the pathological slice image to be processed includes:

[0012] Using a plurality of preset binary classification models to classify the pathological slice images to be processed respectively to obtain a plurality of binary classification results;

[0013] Obtaining an initial tumor classification result of the pathological slice image to be processed according to the multiple binary classification results;

[0014] Among them, each of the binary classification models is used to identify whether the pathological slice image to be processed is a corresponding preset tumor category. When all of the binary classification models identify that the pathological slice image to be processed is not a corresponding preset tumor category, the pathological slice image to be processed is classified separately.

[0015] In a preferred embodiment of the present invention, before classifying the pathological slice image to be processed, the method further includes:

[0016] It is detected whether the pathological slice image to be processed is a preset meaningless slice image. If not, the pathological slice image to be processed is classified and processed.

[0017] In a preferred embodiment of the present invention, the meaningless slice image includes at least one of a tissue-free region slice image, a bleeding region slice image, and a surgical burn region slice image.

[0018] In a preferred embodiment of the present invention, the step of identifying whether there is hyperplasia and necrosis in the pathological slice image to be processed includes:

[0019] Through a preset classification algorithm, it is identified whether there is hyperplasia and necrosis in the pathological slice image to be processed.

[0020] In a second aspect, the present invention provides a pathological image processing method, comprising:

[0021] Acquire a plurality of pathological slice images, wherein the plurality of pathological slice images are obtained by respectively scanning different positions of the same pathological slice;

[0022] Processing each of the pathological slice images respectively using any of the aforementioned methods to obtain a target tumor classification result for each of the pathological slice images;

[0023] The target tumor classification results of each of the pathological slice images are summarized to obtain the tumor classification results of the pathological slices.

[0024] In a preferred embodiment of the present invention, when there is hyperplasia and necrosis in each of the pathological slice images to be processed, the method further comprises:

[0025] A heat map corresponding to the pathological slice is generated according to each of the pathological slice images to be processed, so as to obtain the location of hyperplasia and necrosis in the pathological slice according to the heat map.

[0026] In a preferred embodiment of the present invention, the method further comprises: performing statistics on the target tumor classification results of each of the pathological slice images to obtain statistical results.

[0027] According to a third aspect, an electronic device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements any of the above-described methods when executing the computer program.

[0028] In a fourth aspect, a computer-readable storage medium is provided, on which a computer program is stored, and when the program is executed by a processor, the steps of any of the above methods are implemented.

[0029] By adopting the above technical solution, the present invention has the following beneficial effects:

[0030] The present invention first classifies the pathological slice image to be processed to obtain an initial tumor classification result of the pathological slice image to be processed; then identifies whether there is hyperplasia and necrosis in the pathological slice image to be processed to obtain a hyperplasia and necrosis recognition result of the pathological slice image to be processed; finally, based on the initial tumor classification result and the hyperplasia and necrosis recognition result, obtains a target tumor classification result of the pathological slice image to be processed, thereby automatically identifying the tumor category of the pathological slice image to be processed, saving time and effort. At the same time, by further classifying in combination with the initial tumor classification result and the hyperplasia and necrosis recognition result, a more accurate target tumor classification result can be obtained. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 Flow chart of the pathological slice image processing method according to Embodiment 1 of the present invention;

[0032] Figure 2 Schematic diagram of the pathological slice image processing method of embodiment 1 of the present invention;

[0033] Figure 3 is a flow chart of a pathological image processing method according to Embodiment 2 of the present invention;

[0034] Figure 4 This is an example diagram of pathological sections in Embodiment 2 of the present invention;

[0035] Figure 5 for Figure 4 Cell morphology distribution diagram corresponding to the pathological section;

[0036] Figure 6 for Figure 4 Gradient class activation map corresponding to the pathological slice;

[0037] Figure 7 is a structural block diagram of a pathological slice image processing system according to Embodiment 3 of the present invention;

[0038] Figure 8 is a structural block diagram of a pathological image processing system according to Embodiment 4 of the present invention;

[0039] Fig. 9 This is a hardware architecture diagram of an electronic device according to Embodiment 5 of the present invention. DETAILED DESCRIPTION

[0040] In order to make the purpose, technical scheme and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0041] The terms used in the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present disclosure. The singular forms "a", "the" and "the" used in the present disclosure and the appended claims are also intended to include plural forms unless the context clearly indicates other meanings. It should also be understood that the term "and / or" used herein refers to and includes any or all possible combinations of one or more associated listed items.

[0042] Example 1

[0043] This embodiment provides a pathological slice image processing method, such as Figure 1 and 2 As shown, the method specifically comprises the following steps:

[0044] S11, obtaining the pathological slice image to be processed.

[0045] In this embodiment, the pathological section image refers to a single image obtained by scanning the pathological section (such as zigzag scanning) using a microscope starting from a set area and with a specified resolution (such as a 40x objective lens and a 10x eyepiece). The pathological section refers to a section made by a pathological histological method from a diseased tissue of a certain size, and the types of pathological sections include but are not limited to HE (hematoxylin-eosin) stained sections, cytopathological sections, etc.

[0046] S12, classifying the pathological slice image to be processed to obtain an initial tumor classification result of the pathological slice image to be processed.

[0047] In this embodiment, the pathological slice images to be processed can be divided into the following seven categories according to the different cell morphologies of the pathological slices: anaplastic astrocytoma, ependymoma, astrocytoma, oligodendrocytoma, inflammation, lymphoma and metastasis. Figure 2 Among them, astrocytoma, ependymoma, astrocytoma, oligodendrocytoma belong to non-gliomas, while inflammation, lymphoma and metastatic tumor belong to gliomas.

[0048] Preferably, this embodiment does not use a conventional multi-classification classifier to perform multi-classification processing on the slice image to be processed, but uses multiple binary classification classifiers to classify the slice image to be processed respectively. The specific process is as follows:

[0049] First, a plurality of preset binary classification models are used to classify the cell morphology of the pathological slice image to be processed, and a plurality of binary classification results are obtained; then, based on the plurality of binary classification results, an initial tumor classification result of the pathological slice image to be processed is obtained. Each of the binary classification models is used to identify whether the pathological slice image to be processed is a corresponding preset tumor category. When all the binary classification models identify that the pathological slice image to be processed is not a corresponding preset tumor category, the pathological slice image to be processed is classified separately.

[0050] Specifically, seven binary classifiers can be used to classify the cell morphology of the pathological slice image to be processed, and each binary classifier is used to distinguish anaplastic astrocytoma / other, ependymoma / other, astrocytoma / other, oligodendrocytoma / other, inflammation / other, lymphoma / other and metastasis / other. Among them, each binary classifier can be implemented by DenseNet, and can also be replaced by other classification algorithms, including but not limited to ResNet, VGG and other deep learning algorithms.

[0051] For example, when the result output by a binary classifier is not "other", and the results output by several other classifiers are all "other", the output result of the binary classifier is used as the initial tumor classification result of the pathological section image to be processed.

[0052] When the results output by multiple binary classifiers are all "other", the pathological section image to be processed is identified as an unknown category.

[0053] This embodiment decomposes the multi-classification problem into multiple binary classification problems, which not only reduces the learning difficulty of the classifier and improves the training efficiency of the classifier, but also helps to identify rare unknown cases that do not appear in the training set by separately classifying the pathological section images to be processed as unknown categories when the results output by multiple binary classifiers are all "others".

[0054] S13, identifying whether there is hyperplasia and necrosis in the pathological slice image to be processed, and obtaining a hyperplasia and necrosis identification result of the pathological slice image to be processed.

[0055] Hyperplasia and necrosis are important image features for identifying tumor categories, and there is a large amount of hyperplasia and necrosis in high-risk glioma pathological sections.

[0056] In order to identify whether there is hyperplasia and necrosis in the image, a hyperplasia and necrosis detection model can be trained for detection. However, there are two difficulties in training the hyperplasia and necrosis detection model: first, hyperplasia and necrotic tissue exist in a large area on a single slice image, and it is difficult to mark an accurate detection frame for training; second, for some slices with obvious high-risk features, the tumor cell nuclear density is very high and the color is very dark, which makes it difficult to see the hyperplasia and necrosis on the image, so there will be missed labels.

[0057] In this regard, this embodiment preferably uses a preset classification algorithm instead of a detection model to identify whether there is hyperplasia and necrosis in the pathological slice image to be processed. The classification algorithm here can be implemented using DenseNet, or can be replaced by other classification algorithms, including but not limited to ResNet, VGG and other deep learning algorithms.

[0058] S14, obtaining a target tumor classification result of the pathological slice image to be processed according to the initial tumor classification result and the hyperplasia and necrosis identification result.

[0059] Specifically, this embodiment further classifies the tumor based on the initial tumor classification results and the hyperplasia and necrosis recognition results. Figure 2 As shown, when the initial tumor classification result is ependymoma and there is hyperplasia and necrosis in the pathological slice image to be processed, the target tumor classification result of the pathological slice image to be processed is identified as anaplastic ependymoma; when the initial tumor classification result is astrocytoma and there is hyperplasia and necrosis in the pathological slice image to be processed, the target tumor classification result of the pathological slice image to be processed is identified as glioblastoma; when the initial tumor classification result is oligodendrocytoma and there is hyperplasia and necrosis in the pathological slice image to be processed, the target tumor classification result of the pathological slice image to be processed is identified as anaplastic oligodendrocytoma. For other initial tumor classification results except ependymoma, glioblastoma and anaplastic oligodendrocytoma, they can be directly used as the corresponding target tumor classification results.

[0060] Preferably, before executing step S12, this embodiment further includes: detecting whether the pathological slice image to be processed is a preset meaningless slice image, if so, ending the process, if not, executing step S12 to classify the pathological slice image to be processed.

[0061] In this embodiment, the meaningless slice images include at least one of slice images of a tissue-free region, a bleeding region, and a surgical scald region. The images of these three regions have no reference value for slice diagnosis.

[0062] There is no nuclear structure in the disorganized area and the hemorrhage area. For the HE slice image, since the HE stained slice dyes the cell nucleus purple, the red blood cells in the hemorrhage area are red, and the disorganized area is gray, the disorganized area slice image and the hemorrhage area slice image can be distinguished based on the simple RGB color ratio.

[0063] There are a lot of hyperplastic tissues in the surgical burn area, which mostly appear at the edge of the slice, so the position information and image information of the slice image can be input into the pre-trained classification algorithm for identification. The classification algorithm here can be implemented using DenseNet, or it can be replaced by other classification algorithms, including but not limited to ResNet, VGG and other deep learning algorithms.

[0064] This embodiment can achieve rapid screening of meaningless slice images by detecting whether the pathological slice image to be processed is a preset meaningless slice image before classification.

[0065] In summary, this embodiment first classifies the pathological slice image to be processed to obtain the initial tumor classification result of the pathological slice image to be processed; then identifies whether there is hyperplasia and necrosis in the pathological slice image to be processed to obtain the hyperplasia and necrosis identification result of the pathological slice image to be processed; finally, based on the initial tumor classification result and the hyperplasia and necrosis identification result, the target tumor classification result of the pathological slice image to be processed is obtained, thereby automatically identifying the tumor category of the pathological slice image to be processed, saving time and effort. At the same time, by combining the initial tumor classification result and the hyperplasia and necrosis identification result for classification, a more accurate target tumor classification result can be obtained.

[0066] Example 2

[0067] This embodiment provides a pathological image processing method, such as Figure 6 As shown, the method specifically comprises the following steps:

[0068] S21, acquiring a plurality of pathological slice images, wherein the plurality of pathological slice images are obtained by respectively scanning different positions of the same pathological slice.

[0069] For example, the plurality of pathological slice images refer to all images obtained by scanning (eg, zigzag scanning) the pathological slices starting from a set area using a microscope with a specified resolution (eg, 40x objective lens, 10x eyepiece).

[0070] S22, processing each of the pathological slice images respectively using the method described in any of the aforementioned embodiments to obtain a target tumor classification result for each of the pathological slice images.

[0071] Specifically, for each pathological slice image, the pathological slice image is processed respectively using the method described in any of the aforementioned embodiments to obtain a target tumor classification result of the pathological slice image.

[0072] S23, summarizing the target tumor classification results of each of the pathological slice images to obtain the tumor classification results of the pathological slices.

[0073] In this embodiment, a multi-instance learning algorithm may be specifically used to summarize and learn the target tumor classification results of each pathological slice image to obtain the final tumor classification result of the corresponding pathological slice.

[0074] For example, when the target tumor classification results of a single pathology section image are divided into the following 12 categories: meaningless staining, unknown, metastasis, lymphoma, inflammation, oligodendrocytoma, anaplastic oligodendrocytoma, astrocytoma, anaplastic astrocytoma, glioblastoma, ependymoma and anaplastic ependymoma, the corresponding pathology sections can be divided into the following 11 categories: unknown, metastasis, lymphoma, inflammation, oligodendrocytoma, anaplastic oligodendrocytoma, astrocytoma, anaplastic astrocytoma, glioblastoma, ependymoma and anaplastic ependymoma.

[0075] This embodiment uses the method described in any of the aforementioned embodiments to process the pathological slice images at different positions of the same pathological slice respectively to obtain the target tumor classification results of each pathological slice image, and then obtains the tumor classification results of the pathological slice based on the target tumor classification results of each pathological slice image, thereby realizing automatic diagnosis of the pathological slice.

[0076] Preferably, this embodiment can also generate a corresponding cell morphology distribution map according to the target tumor classification result of each slice image to be processed, so as to provide a reference for doctors in the post-processing stage. Among them, each cell morphology in the cell morphology distribution map corresponds to a target tumor classification result. For example, when the pathological slice is as follows Figure 3 When shown, the corresponding cell morphology distribution diagram is generated as shown in Figure 4 shown.

[0077] Preferably, when there is hyperplasia and necrosis in the pathological slice image to be processed, this embodiment may further include: generating a heat map corresponding to the pathological slice according to each of the pathological slice images to be processed, so as to obtain the location of hyperplasia and necrosis in the pathological slice according to the heat map. The heat map here is preferably as follows Figure 5The gradient class activation map shown can display the specific location of hyperplasia and necrosis in the pathological slice image, which is convenient for locating high-risk areas and helps the pathologist's subsequent diagnosis. The specific process of generating the gradient class activation map in this embodiment is as follows: first, gradient class activation processing is performed on each pathological slice image to be processed, and then a gradient class activation map is generated according to the gradient class activation processing results of all the slice images to be processed.

[0078] Preferably, this embodiment further comprises collecting statistics on the target tumor classification results of each of the pathological slice images to obtain statistical results, which are preferably displayed in a pie chart or the like for reference by doctors in the post-processing stage.

[0079] Example 3

[0080] This embodiment provides a pathological slice image processing system. Figure 7 As shown, the system includes:

[0081] A first image acquisition module 11, used to acquire a pathological slice image to be processed;

[0082] An initial classification module 12 is used to classify the pathological slice image to be processed to obtain an initial tumor classification result of the pathological slice image to be processed;

[0083] A hyperplasia identification module 13 is used to identify whether there is hyperplasia and necrosis in the pathological slice image to be processed, and obtain a hyperplasia and necrosis identification result of the pathological slice image to be processed;

[0084] The target classification module 14 is used to obtain the target tumor classification result of the pathological slice image to be processed according to the initial tumor classification result and the hyperplasia and necrosis identification result.

[0085] Preferably, the initial classification module 12 is specifically used for:

[0086] Using a plurality of preset binary classification models to classify the pathological slice images to be processed respectively to obtain a plurality of binary classification results;

[0087] An initial tumor classification result of the pathological slice image to be processed is obtained according to the multiple binary classification results.

[0088] Preferably, the system also includes: an image detection module, which is used to detect whether the pathological slice image to be processed is a preset meaningless slice image before the initial classification module 12 classifies the pathological slice image to be processed. If not, calling the initial classification module 12 to classify the pathological slice image to be processed.

[0089] Preferably, the meaningless slice image includes at least one of a tissue-free region slice image, a bleeding region slice image and a surgical burn region slice image.

[0090] Preferably, the hyperplasia identification module 13 is specifically used for:

[0091] Through a preset classification algorithm, it is identified whether there is hyperplasia and necrosis in the pathological slice image to be processed.

[0092] This embodiment obtains an initial tumor classification result of the pathological slice image to be processed by classifying the pathological slice image to be processed; identifies whether there is hyperplasia and necrosis in the pathological slice image to be processed, and obtains a hyperplasia and necrosis recognition result of the pathological slice image to be processed; obtains a target tumor classification result of the pathological slice image to be processed based on the initial tumor classification result and the hyperplasia and necrosis recognition result, thereby automatically identifying the tumor category of the pathological slice image to be processed, saving time and effort. At the same time, by combining the initial tumor classification result and the hyperplasia and necrosis recognition result for classification, a more accurate target tumor classification result can be obtained.

[0093] Example 4

[0094] This embodiment provides a pathological image processing system. Figure 8 As shown, the system includes:

[0095] The second image acquisition module 21 is used to acquire a plurality of pathological slice images, wherein the plurality of pathological slice images are obtained by respectively scanning different positions of the same pathological slice;

[0096] An image processing module 22, configured to process each of the pathological slice images respectively using the method described in any of the above embodiments to obtain a target tumor classification result of each of the pathological slice images;

[0097] The slice classification module 23 is used to summarize the target tumor classification results of each of the pathological slice images to obtain the tumor classification results of the pathological slices.

[0098] This embodiment uses the method described in any of the aforementioned embodiments to process the pathological slice images at different positions of the same pathological slice respectively, to obtain the target tumor classification result of each pathological slice image, and based on the target tumor classification result of each pathological slice image, obtain the tumor classification result of the pathological slice, so as to realize automatic diagnosis of the pathological slice.

[0099] Preferably, when there is hyperplasia and necrosis in each of the pathological slice images to be processed, the method further comprises:

[0100] A heat map corresponding to the pathological slice is generated according to each of the pathological slice images to be processed, so as to obtain the location of hyperplasia and necrosis in the pathological slice according to the heat map, and the heat map is preferably a gradient activation map.

[0101] Preferably, the method further comprises: generating a pie chart according to the target tumor classification result of each of the pathological slice images.

[0102] As for the system embodiment, since it basically corresponds to the method embodiment, the relevant parts can refer to the partial description of the method embodiment. The system embodiment described above is only illustrative, wherein the units described as separate components may or may not be physically separated, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of the present invention. Ordinary technicians in this field can understand and implement it without paying creative work.

[0103] Example 5

[0104] Fig. 9 FIG. 1 is a schematic diagram of the structure of an electronic device according to an exemplary embodiment of the present invention, and is a block diagram of an exemplary electronic device 60 suitable for implementing the embodiments of the present invention. Fig. 9 The electronic device 60 shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0105] like Figure 6 As shown, the electronic device 60 may be in the form of a general-purpose computing device, for example, it may be a server device. The components of the electronic device 60 may include, but are not limited to: at least one processor 61, at least one memory 62, and a bus 63 connecting different system components (including the memory 62 and the processor 61).

[0106] The bus 63 includes a data bus, an address bus, and a control bus.

[0107] The memory 62 may include a volatile memory, such as a random access memory (RAM) 621 and / or a cache memory 622 , and may further include a read-only memory (ROM) 623 .

[0108] The memory 62 may also include a program tool 625 (or utility) having a set (at least one) of program modules 624, such program modules 624 including but not limited to: an operating system, one or more application programs, other program modules, and program data, each of which or some combination may include an implementation of a network environment.

[0109] The processor 61 executes various functional applications and data processing by running the computer program stored in the memory 62, such as the method provided in any of the above embodiments.

[0110] The electronic device 60 may also communicate with one or more external devices 64 (e.g., keyboards, pointing devices, etc.). Such communication may be performed via an input / output (I / O) interface 65. Furthermore, the model-generated electronic device 60 may also communicate with one or more networks (e.g., a local area network (LAN), a wide area network (WAN), and / or a public network, such as the Internet) via a network adapter 66. As shown, the network adapter 66 communicates with other modules of the model-generated electronic device 60 via a bus 63. It should be understood that, although not shown in the figure, other hardware and / or software modules may be used in conjunction with the model-generated electronic device 60, including but not limited to: microcode, device drivers, redundant processors, external disk drive arrays, RAID (disk array) systems, tape drives, and data backup storage systems, etc.

[0111] It should be noted that although several units / modules or sub-units / modules of the electronic device are mentioned in the above detailed description, this division is merely exemplary and not mandatory. In fact, according to an embodiment of the present invention, the features and functions of two or more units / modules described above can be embodied in one unit / module. Conversely, the features and functions of one unit / module described above can be further divided into multiple units / modules to be embodied.

[0112] Example 6

[0113] This embodiment provides a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the steps of the method provided in any of the above embodiments are implemented.

[0114] The readable storage medium may include but is not limited to: a portable disk, a hard disk, a random access memory, a read-only memory, an erasable programmable read-only memory, an optical storage device, a magnetic storage device or any suitable combination of the above.

[0115] In a possible implementation manner, the present disclosure may also be implemented in the form of a program product, which includes program code. When the program product is run on a terminal device, the program code is used to enable the terminal device to execute the steps of implementing the method provided in any of the above embodiments.

[0116] Among them, the program code for executing the present disclosure can be written in any combination of one or more programming languages, and the program code can be executed completely on the user device, partially on the user device, as an independent software package, partially on the user device and partially on a remote device, or completely on the remote device.

[0117] Although the specific embodiments of the present disclosure are described above, those skilled in the art should understand that this is only an example, and the protection scope of the present disclosure is defined by the appended claims. Those skilled in the art may make various changes or modifications to these embodiments without departing from the principles and essence of the present disclosure, but these changes and modifications all fall within the protection scope of the present disclosure.

Claims

1. A pathological slice image processing method, characterized in that: include: Acquire the pathological slice image to be processed; Performing classification processing on the pathological slice image to be processed to obtain an initial tumor classification result of the pathological slice image to be processed; Identify whether there is hyperplasia and necrosis in the pathological slice image to be processed, and obtain a hyperplasia and necrosis identification result of the pathological slice image to be processed; According to the initial tumor classification result and the hyperplasia and necrosis identification result, obtaining a target tumor classification result of the pathological slice image to be processed; The classifying process of the pathological slice image to be processed to obtain an initial tumor classification result of the pathological slice image to be processed includes: Using a plurality of preset binary classification models to classify the pathological slice image to be processed, respectively, to obtain a plurality of binary classification results, wherein each of the binary classification models is used to identify whether the pathological slice image to be processed is a corresponding preset tumor category; Obtaining an initial tumor classification result of the pathological slice image to be processed according to the multiple binary classification results; When all the binary classification models identify that the pathological slice image to be processed does not belong to the corresponding preset tumor category, the pathological slice image to be processed is classified separately.

2. The pathological slice image processing method according to claim 1, characterized in that: Before classifying the pathological slice image to be processed, the method further includes: It is detected whether the pathological slice image to be processed is a preset meaningless slice image. If not, the pathological slice image to be processed is classified and processed.

3. The pathological section image processing method according to claim 2, characterized in that: The meaningless slice image includes at least one of a tissue-free region slice image, a bleeding region slice image, and a surgical scald region slice image.

4. The pathological slice image processing method according to claim 1, characterized in that: The identifying whether there is hyperplasia and necrosis in the pathological slice image to be processed includes: Through a preset classification algorithm, it is identified whether there is hyperplasia and necrosis in the pathological slice image to be processed.

5. A pathological image processing method, characterized in that: include: Acquire a plurality of pathological slice images, wherein the plurality of pathological slice images are obtained by respectively scanning different positions of the same pathological slice; Processing each of the pathological slice images respectively using the method described in any one of claims 1 to 4 to obtain a target tumor classification result for each of the pathological slice images; The target tumor classification results of each of the pathological slice images are summarized to obtain the tumor classification results of the pathological slices.

6. The pathological image processing method according to claim 5, characterized in that: When there is hyperplasia and necrosis in the pathological slice image to be processed, the method further includes: A heat map corresponding to the pathological slice is generated according to each of the pathological slice images to be processed, so as to obtain the location of hyperplasia and necrosis in the pathological slice according to the heat map.

7. The pathological image processing method according to claim 6, characterized in that: The method further comprises: performing statistics on the target tumor classification results of each of the pathological slice images to obtain statistical results.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the method according to any one of claims 1 to 7 is implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images