Medical dyeing image processing method and device, storage medium and electronic equipment

Abstract

The invention provides a medical dyeing image processing method and device, a storage medium and electronic equipment. The medical dyeing image processing method comprises the following steps: acquiring a to-be-processed image, wherein the to-be-processed image is a medical dyeing image; and if the to-be-processed image has the dyeing problem, processing the to-be-processed image by using a non-paired image migration model so as to correct the dyeing problem of the to-be-processed image. According to the medical dyeing image processing method, dyeing deviation and / or dyeing abnormity in the medical dyeing image can be corrected.

Description

Technical Field

[0001] The present application belongs to the technical field of medical image processing, and relates to a medical image processing method, and in particular to a medical staining image processing method, device, storage medium and electronic equipment. Background Art

[0002] Medical staining images are images obtained by scanning cells or tissue samples after staining. Staining technology plays an important role in medical diagnosis and research. It can highlight the specific structure, components or characteristics of cells and tissues, thereby helping doctors and researchers to accurately locate and identify them. Medical staining images can use a variety of different staining methods and stains, and the specific choice depends on the object of study or the part being examined. For example, in the TCT (ThinPrep Cytology Test) liquid-based thin layer cell test, medical staff usually use Papanicolaou stain to stain cells or tissues for cervical cancer screening. However, due to the influence of various factors such as sample preparation, stains and scanning equipment, the final medical staining images often have problems such as staining deviation or staining abnormality, which will have an adverse effect on computer-assisted diagnosis. Summary of the invention

[0003] The embodiments of the present application provide a medical staining image processing method, device, storage medium and electronic device for solving problems such as staining deviation or staining abnormality.

[0004] In a first aspect, an embodiment of the present application provides a medical staining image processing method, which includes: obtaining an image to be processed, wherein the image to be processed is a medical staining image; if there is a staining problem in the image to be processed, processing the image to be processed using an unpaired image migration model to correct the staining problem of the image to be processed.

[0005] In an implementation of the first aspect, the medical staining image processing method further includes: acquiring a color distance between the image to be processed and a standard image; and judging whether the image to be processed has a staining problem according to the color distance.

[0006] In an implementation of the first aspect, obtaining the color distance between the image to be processed and the standard image includes: performing clustering processing on the image to be processed to obtain the cluster center points of the image to be processed, the cluster center points including background pixel cluster center points, pre-keratinized cytoplasm pixel cluster center points and cornified cytoplasm cluster center points; obtaining the color distance based on the distance between the cluster center points of the image to be processed and the cluster center points of the standard image.

[0007] In an implementation of the first aspect, the training method of the image migration model includes: acquiring multiple training images; dividing the training images into a first image group and a second image group according to the color distance between the training images and the standard images, wherein the first image group is the standard group; and training the image migration model using the first image group and the second image group.

[0008] In an implementation manner of the first aspect, the image migration model is a cycle consistency generative adversarial network, and the loss function used in the training of the image migration model includes adversarial loss, cycle consistency loss and structural consistency loss.

[0009] In an implementation of the first aspect, the training image is an image obtained after full image cutting, and the training method of the image migration model also includes: obtaining the proportion of the background area in each of the training images; and screening the training images according to the proportion.

[0010] In an implementation of the first aspect, the medical staining image is a TCT Papanicolaou staining image.

[0011] In a second aspect, an embodiment of the present application provides a medical staining image processing device, which includes: an image acquisition module, used to acquire an image to be processed, wherein the image to be processed is a medical staining image; an image processing module, used to process the image to be processed using an unpaired image migration model when there is a staining problem in the image to be processed, so as to correct the staining problem of the image to be processed.

[0012] In a third aspect, an embodiment of the present application provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the method described in any one of the first aspects of the present application is implemented.

[0013] In a fourth aspect, an embodiment of the present application provides an electronic device, comprising: a memory storing a computer program; and a processor communicatively connected to the memory, for executing any of the methods described in the first aspect of the present application when the computer program is called.

[0014] The medical staining image processing method provided in the embodiment of the present application can use the unpaired image migration model to correct the staining problem of the image to be processed. In this way, the impact of inconsistent pathological staining problems on the artificial intelligence model can be reduced, thereby improving the stability and accuracy of artificial intelligence-assisted diagnosis in TCT detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 Shown is a schematic diagram of an application scenario of the medical staining image processing method provided in an embodiment of the present application.

[0016] Figure 2 Shown is a flow chart of the medical staining image processing method provided in an embodiment of the present application.

[0017] Figure 3 Shown is a flow chart for determining whether there is a staining problem in an embodiment of the present application.

[0018] Figure 4 Shown is a flow chart of obtaining color distance in an embodiment of the present application.

[0019] Figure 5 Shown is a flowchart for training an image migration model in an embodiment of the present application.

[0020] Figure 6 Shown is a flowchart for screening training images in an embodiment of the present application.

[0021] Figure 7 Shown is a structural schematic diagram of the image migration model in an embodiment of the present application.

[0022] Figure 8 Shown is a schematic diagram of the structure of the medical staining image processing device in an embodiment of the present application.

[0023] Component number description

[0024] 100 electronic devices

[0025] 101, 107 processors

[0026] 102 Output Devices

[0027] 103 Input Devices

[0028] 104 Memory Units

[0029] 105 Communication interface

[0030] 106 Storage Media

[0031] 800 Medical staining image processing device

[0032] 810 Image acquisition module

[0033] 820 Image Processing Module

[0034] Steps S21 to S22

[0035] Steps S31-32

[0036] Steps S41 to S42

[0037] Steps S51 to S53

[0038] Steps S61 to S62 DETAILED DESCRIPTION

[0039] The following describes the embodiments of the present application through specific examples, and those skilled in the art can easily understand other advantages and effects of the present application from the contents disclosed in this specification. The present application can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present application. It should be noted that the following embodiments and features in the embodiments can be combined with each other without conflict.

[0040] It should be noted that the illustrations provided in the following embodiments are only used to illustrate the basic concept of the present application in a schematic manner, and therefore the illustrations only show components related to the present application rather than being drawn according to the number, shape and size of components in actual implementation. In actual implementation, the type, quantity and proportion of each component may be changed arbitrarily, and the component layout may also be more complicated.

[0041] Staining technology plays an important role in medical diagnosis and research. For example, in TCT liquid-based thin-layer cell testing, cells or tissues are stained with Pap stain to screen for cervical cancer. However, differences in the operating specifications of different scanning devices and different inspectors may cause the final medical pathology images to have staining differences. Specifically, differences in scanning equipment may cause slight color differences and sensitivities in the same hue (such as blue and pink in TCT Pap staining), which is a staining deviation. Differences in the operating specifications of different inspectors may cause completely different tones in the imaging (such as orange and yellow in TCT Pap staining), which is a staining abnormality. When the above medical pathology images are applied to artificial intelligence (AI)-assisted diagnosis algorithms, the stability and conclusion accuracy of the AI-assisted diagnosis algorithm will be reduced. In some technical solutions, a method of re-coloring based on the brightness information (grayscale image) of the original stained pathological image and the information of the cytoplasmic position of pre-keratinized cells (mask image) is used to normalize the color gamut of a set of stained images. On the one hand, these technical solutions cannot be used to process pathological images with abnormal staining. On the other hand, pathology slides are often mixed with blood, lint, and impurities. The recoloring methods used in these technical solutions cannot accurately restore the colors of these situations, resulting in a high false positive rate.

[0042] At least for the above problems, the embodiment of the present application provides a medical staining image processing method. The method can use an unpaired image migration model to correct the staining problem of the image to be processed. In this way, the impact of inconsistent pathological staining problems on the AI-assisted diagnosis algorithm can be reduced, thereby improving the stability and accuracy of the AI-assisted diagnosis algorithm in TCT detection.

[0043] The technical solutions in the embodiments of the present application will be described in detail below in conjunction with the drawings in the embodiments of the present application.

[0044] The medical staining image processing method provided in the embodiment of the present application can be applied to electronic devices. Figure 1 The structure diagram of the electronic device 100 in the embodiment of the present application is shown. Figure 1 As shown, the electronic device 100 includes a processor 101 connected to one or more data storage units. The data storage unit may include a storage medium 106 and a memory unit 104. The storage medium 106 may be read-only, such as a read-only memory (ROM), or readable and writable, such as a hard disk or a flash memory. The memory unit 104 may be a random access memory (RAM). The memory unit 104 may be integrated with the processor 101 or may be an independent component. The processor 101 is the control center of the electronic device 100, and is used to execute program codes to implement functions corresponding to the program instructions. Optionally, the processor 101 includes one or more central processing units (CPU), such as Figure 1 Optionally, the electronic device 100 includes more than one processor, for example, Figure 1 The processors 101 and 107 are shown. Both the processors 101 and 107 can be single-core processors or multi-core processors. As used herein, the term "processor" refers to one or more devices, circuits and / or processing cores for processing data such as computer program instructions.

[0045] The CPU of processor 101 and / or 107 stores the executed program code in memory unit 104 or storage medium 106. Optionally, the program code stored in storage medium 106 can be copied to memory unit 104 for execution by the processor. The processor can control the operation of electronic device 100 by controlling the execution of other programs, controlling the communication with peripheral devices, and controlling the use of resources of electronic device 100 through the kernel.

[0046] The electronic device 100 may further include a communication interface 105 , through which the electronic device 100 may communicate with another device or system directly or through an external network.

[0047] Optionally, the electronic device 100 also includes an output device 102 and an input device 103. The output device 102 is connected to the processor 101, and can display output information in one or more ways. An example of the output device 102 is a visual display device, for example, a liquid crystal display (LCD), a light-emitting diode (LED) display, a cathode ray tube (CRT) or a projector. The input device 103 is connected to the processor 101, and can receive the user's input in one or more ways. The example of the input device 103 includes a mouse, a keyboard, a touch screen device, a sensing device, etc.

[0048] The above elements of the electronic device 100 may be connected to each other via any one or more combinations of buses such as a data bus, an address bus, a control bus, an extension bus, and a local bus.

[0049] The electronic device 100 may be a general electronic device or an application-specific electronic device. As a practical example, the electronic device 100 may be a storage array, an application server, a supercomputer, a desktop computer, a laptop computer, a personal digital assistant (PDA), a mobile phone, a tablet computer, a wireless terminal device, a telecommunication device, or a computer having a plurality of such devices. Figure 1 Any other device with similar structure as shown. However, the present application is not limited to any specific type of electronic device. The program codes with different functions stored in the memory 104 are executed by the processor (processor 101 or processor 107) to form a process. When the process is running, the processor needs to allocate a memory space to each process to store data generated during the process running.

[0050] Figure 2 The flowchart of the medical staining image processing method provided by the embodiment of the present application is shown. Figure 1 The processor 101 and / or 107 shown is implemented. Figure 2 As shown, the medical staining image processing method provided in the embodiment of the present application may include the following steps S21 and S22.

[0051] S21, obtaining an image to be processed, where the image to be processed is a medical staining image, such as a TCT Papanicolaou staining image.

[0052] S22, if there is a staining problem in the image to be processed, the image to be processed is processed using an unpaired image transfer model to correct the staining problem of the image to be processed. An unpaired image transfer model refers to a model that does not require input of paired data between image pairs for training in an image transfer task. The unpaired image transfer model can perform image conversion in the absence of paired data by learning the mapping relationship between the source domain and the target domain. The staining problem of the image to be processed includes but is not limited to staining deviation and / or staining anomaly.

[0053] See also Figure 3 In one embodiment of the present application, the medical staining image processing method may further include the following steps S31 and S32.

[0054] S31, obtaining the color distance between the image to be processed and the standard image, wherein the standard image is a pathological image with normal staining.

[0055] S32, judging whether the image to be processed has a staining problem according to the color distance. In some implementations, the relationship between the color distance and the distance threshold can be used to judge whether the image to be processed has a staining problem, wherein the distance threshold can be configured according to actual needs. Specifically, if the color distance between the image to be processed and the standard image is less than the distance threshold, it is judged that the image to be processed has no staining problem, otherwise it is judged that the image to be processed has a staining problem and needs to be corrected.

[0056] Figure 4 The flowchart of obtaining the color distance between the image to be processed and the standard image in the embodiment of the present application is shown. Figure 4 As shown, in the embodiment of the present application, obtaining the color distance between the image to be processed and the standard image includes the following steps S41 and S42.

[0057] S41, clustering the image to be processed to obtain the cluster center point of the image to be processed. Wherein, the cluster center point includes the background pixel cluster center point, the pre-keratinization cytoplasm pixel cluster center point and the cornified cytoplasm cluster center point. Specifically, in the embodiment of the present application, the image to be processed can be clustered at the pixel level to cluster the pixels in the image to be processed into three categories, namely, background pixel points, pre-keratinization cytoplasm pixel points and cornified cytoplasm pixel points. According to the clustering result, the background pixel cluster center point, the pre-keratinization cytoplasm pixel cluster center point and the cornified cytoplasm cluster center point can be obtained. In addition, in the embodiment of the present application, the above-mentioned similar method can be used to obtain the background pixel cluster center point, the pre-keratinization cytoplasm pixel cluster center point and the cornified cytoplasm cluster center point of the standard image.

[0058] In some implementations, k-means or spectral clustering can be used to cluster pixels in the image to be processed into three colors, such as white background, blue-green prekeratinized cytoplasm, and pink keratinized cytoplasm.

[0059] S42, obtaining a color distance according to the distance between the cluster center point of the image to be processed and the cluster center point of the standard image.

[0060] In some implementations, the color distance may be quantified by the sum of the shortest Euclidean distances between the three cluster center points of the image to be processed and the three cluster center points of the standard image, but the present application is not limited thereto.

[0061] Figure 5 The flowchart of the training method of the image migration model in the embodiment of the present application is shown. Figure 5 As shown, the training method of the image migration model in the embodiment of the present application includes the following steps S51 to S53.

[0062] S51, obtaining a plurality of training images, where the training images are medical stained images.

[0063] S52, dividing the training image into a first image group and a second image group according to the color distance between the training image and the standard image, wherein the first image group is the standard group.

[0064] In some implementations, the method for dividing the training images includes: selecting a normally stained image from the training images as a standard image; obtaining the color distance between the remaining training images and the standard image; dividing all the training images whose color distances with the standard image are less than a distance threshold into a first image group, and dividing the remaining images into a second image group. The method for obtaining the color distance between the training image and the standard image is similar to the above steps S41 and S42, and will not be described in detail here.

[0065] S53: Train the unpaired image migration model using the first image group and the second image group.

[0066] In some implementations, the training image is an image obtained by cutting a whole slide imaging (WSI). During the cutting process, a portion of the training image may contain very little cell information or no cell information (only some impurities). At least in view of the above problem, in one embodiment of the present application, the training method of the image migration model also includes Figure 6 Steps S61 and S62 are shown.

[0067] S61, obtaining the ratio of the background area in each training image. In some implementations, the ratio of the background area can be obtained according to the clustering result of the training image, but the present application is not limited thereto.

[0068] S62, screening the training images according to the above ratio. In some implementations, if the ratio of the background area in the training image is greater than a ratio threshold, the training image is excluded from the training set. The ratio threshold can be configured according to actual needs.

[0069] In one embodiment of the present application, the image migration model is a cycle-consistent generative adversarial network (CycleGAN). Figure 7 It is shown as a schematic diagram of the structure of the image migration model in the embodiment of the present application. Among them, Y is the color gamut space represented by the standard group, and X is the color gamut space represented by the remaining training images outside the standard group. The image migration model in the embodiment of the present application includes two generators F and G, and a discriminator D. By adopting the unsupervised image migration method for model training, those images with color deviation / abnormality that are too far away from the standard image can be corrected by the domain migration method.

[0070] In some implementations, the image migration model uses ResNet as the basic architecture of the generator G and PatchGAN as the basic architecture of the discriminator D.

[0071] AI-assisted diagnosis based on medical staining images not only needs to be based on relative color difference information between cells, but also needs to be based on morphological information such as cytoplasm and nucleus to draw conclusions. Therefore, the consistency of the structural and texture morphological information of the medical staining image before and after image migration is crucial for staining correction. To address this problem, an additional loss function is introduced in an embodiment of the present application to enhance the consistency of the structure, brightness, contrast and relative color difference of the medical staining image before and after image migration. Exemplarily, the loss functions used in the training of the image migration model include adversarial loss, cycle consistency loss and structural consistency loss.

[0072] Adversarial loss includes two-way adversarial loss functions GAN(X→Y) and GAN(Y→X). GAN(X→Y) mainly involves the performance of the discriminator D and the performance of the generator G. The first part is used to measure the performance of the discriminator D. Y The ability of the discriminator D Y The ability to correctly classify real data y. y represents samples from the actual dataset Y (that is, the color space represented by the standard group). The second part is used to measure the ability of the generator G, that is, the ability of the generator G to generate samples that can deceive the discriminator D.Y The ability of the generator G to generate fake data. x represents a sample from the dataset X (that is, the color gamut space represented by the remaining training images outside the standard group). G(x) represents the fake data generated by the generator G using x. The goal of the adversarial loss is to minimize the difference between the fake data generated by the generator G and the real data and maximize the discriminator D Y The ability to distinguish between real data and fake data. The adversarial loss GAN(Y→X) in the other direction is similar to GAN(X→Y) and will not be described here.

[0073] In some implementations, the adversarial losses L in the above two directions are GAN (G,D Y ,X,Y) and L GAN (F,D X ,Y,X) can be expressed by the following formula 1 and formula 2 respectively.

[0074]

[0075]

[0076] Among them, P data (y) represents the distribution function of sample y, P represents the expected value of the V function corresponding to sample y. data (x) represents the distribution function of sample x, Represents the expected value of the V function corresponding to sample x.

[0077] Cycle consistency loss is used to ensure that the migration of an image can return to the original image. Specifically, for an image from dataset X, it is migrated to dataset Y and then migrated back to dataset X. Cycle consistency loss is used to ensure that it can be as close to the original image as possible after being migrated back to dataset X.

[0078] In some implementations, the cycle consistency loss L cyc (G, F) can be expressed by the following formula 3.

[0079]

[0080] The structural consistency loss is used to measure the consistency of the L* channel in the L*a*b color space before and after the image migration. The L* channel is used to identify the brightness of the image and can reflect the local texture structure of the image. In the embodiment of the present application, CycleGAN is used to correct the color deviation and / or color abnormality, and the color gamut of the image is migrated while ensuring that the texture structure remains unchanged. By constraining the brightness, contrast, and structure of the L* channel of the image before and after the migration, the structural consistency of the image during the migration process can be emphasized.

[0081] In some implementations, the structural consistency loss can be expressed using the following formula 4.

[0082]

[0083] Among them, A and B represent two image matrices, μ A and μ B Represent the average values ​​of the pixels of A and B images, σ A and σ B Represent the mean and standard deviation of the pixels in images A and B, respectively, AB represents the standard deviation of image A and image B, that is, the square root of the covariance of image A and image B. c1 and c2 are two constants to ensure the stability of calculation. The values ​​of the two can be configured according to actual needs. In the embodiment of the present application, the structural consistency loss consists of two parts. The first part is It is used to measure the structural consistency of the abnormal staining domain to the standard group. It is used to measure the structural consistency of the standard group in the process of returning to the abnormal staining domain. Wherein, L* represents the value of the L* channel.

[0084] In some implementations, the loss function used in the training of the image transfer model can be expressed as follows: Among them, λ1 and λ2 are two hyperparameters used to control the weight of the corresponding loss function in the total loss function. Their values ​​can be configured according to actual needs.

[0085] The protection scope of the medical staining image processing method provided in the embodiment of the present application is not limited to the execution order of the steps listed in this embodiment. All solutions implemented by adding, reducing or replacing steps in the prior art based on the principles of the present application are included in the protection scope of the present application.

[0086] An embodiment of the present application also provides a medical staining image processing device, which can implement the medical staining image processing method described in the present application. However, the implementation device of the medical staining image processing method described in the present application includes but is not limited to the structure of the medical staining image processing device listed in this embodiment. All structural deformations and replacements of the prior art made according to the principles of the present application are included in the protection scope of the present application.

[0087] Figure 8 The structure diagram of the medical staining image processing device 800 in one embodiment of the present application is shown. Figure 8As shown, the medical staining image processing device 800 includes an image acquisition module 810 and an image processing module 820. The image acquisition module 810 is used to acquire an image to be processed, which is a medical staining image. The image processing module 820 is used to process the image to be processed using an unpaired image transfer model when there is a staining problem in the image to be processed, so as to correct the staining problem of the image to be processed.

[0088] It should be noted that the image acquisition module 810 and the image processing module 820 included in the medical staining image processing device 800 are Figure 2 Steps S21 and S22 in the medical staining image processing method shown correspond to each other and will not be described in detail here.

[0089] In the several embodiments provided in the present application, it should be understood that the disclosed system, device or method can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of modules / units is only a logical function division. There may be other division methods in actual implementation, such as multiple modules or units can be combined or integrated into another system, 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 an indirect coupling or communication connection through some interfaces, devices or modules or units, which can be electrical, mechanical or other forms.

[0090] The modules / units described as separate components may or may not be physically separated, and the components displayed as modules / units may or may not be physical modules, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules / units may be selected according to actual needs to achieve the purpose of the embodiments of the present application. For example, the functional modules / units in the various embodiments of the present application may be integrated into one processing module, or each module / unit may exist physically separately, or two or more modules / units may be integrated into one module / unit.

[0091] Those of ordinary skill in the art should further appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in the above description according to function. 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 this application.

[0092] The present application also provides a computer-readable storage medium on which a computer program is stored, and when the computer program is executed, the medical staining image processing method provided by any embodiment of the present application is implemented. A person of ordinary skill in the art can understand that all or part of the steps in the method for implementing the above-mentioned embodiment can be completed by instructing the processor through a program, and the program can be stored in a computer-readable storage medium, and the storage medium is a non-transitory medium, such as a random access memory, a read-only memory, a flash memory, a hard disk, a solid-state hard disk, a magnetic tape, a floppy disk, an optical disc, and any combination thereof. The above-mentioned storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or a data center that includes one or more available media integrated. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a digital video disc (DVD)), or a semiconductor medium (e.g., a solid-state disk (SSD)), etc.

[0093] The embodiment of the present application also provides an electronic device. The electronic device includes a memory and a processor. The memory is used to store a computer program. In some possible implementations, the memory may include: ROM, RAM, a disk, a USB flash drive, a memory card, or a CD, etc., which can store program codes. The processor is communicatively connected to the memory, and is used to execute the computer program stored in the memory, so that the electronic device performs the medical staining image processing method.

[0094] In some possible implementations, the electronic device 800 provided in the embodiment of the present application may further include a display 830. The display 830 is communicatively connected to the memory 810 and the processor 820, and is used to display a graphical user interface (GUI) related to the medical staining image processing method.

[0095] The above embodiments are merely illustrative of the principles and effects of the present application and are not intended to limit the present application. Anyone familiar with the technology may modify or change the above embodiments without violating the spirit and scope of the present application. Therefore, all equivalent modifications or changes made by a person of ordinary skill in the art without departing from the spirit and technical ideas disclosed in the present application shall still be covered by the claims of the present application.

Claims

1. A medical staining image processing method, characterized in that: The medical staining image processing method comprises: Acquire an image to be processed, wherein the image to be processed is a medical staining image; If the image to be processed has a coloring problem, the image to be processed is processed using an unpaired image migration model to correct the coloring problem of the image to be processed.

2. The medical staining image processing method according to claim 1, characterized in that: The medical staining image processing method further comprises: Obtaining the color distance between the image to be processed and the standard image; It is determined whether the image to be processed has a coloring problem according to the color distance.

3. The medical staining image processing method according to claim 2, characterized in that: Obtaining the color distance between the image to be processed and the standard image includes: Performing clustering processing on the image to be processed to obtain cluster center points of the image to be processed, wherein the cluster center points include background pixel cluster center points, pre-keratinized cytoplasm pixel cluster center points and keratinized cytoplasm cluster center points; The color distance is obtained according to the distance between the cluster center point of the image to be processed and the cluster center point of the standard image.

4. The medical staining image processing method according to claim 1, characterized in that: The training method of the image migration model includes: Get multiple training images; Dividing the training images into a first image group and a second image group according to a color distance between the training images and the standard images, wherein the first image group is a standard group; The image migration model is trained using the first image group and the second image group.

5. The medical staining image processing method according to claim 4, characterized in that: The image migration model is a cycle consistency generative adversarial network, and the loss functions used in the training of the image migration model include adversarial loss, cycle consistency loss and structural consistency loss.

6. The medical staining image processing method according to claim 4, characterized in that: The training image is an image obtained after full image cutting, and the training method of the image migration model also includes: Obtaining the ratio of the background area in each of the training images; The training images are screened according to the ratio.

7. The medical staining image processing method according to any one of claims 1 to 6, characterized in that: The medical staining image is a TCT Papanicolaou staining image.

8. A medical staining image processing device, characterized in that: The medical staining image processing device comprises: An image acquisition module, used for acquiring an image to be processed, wherein the image to be processed is a medical staining image; The image processing module is used to process the image to be processed by using an unpaired image migration model when the image to be processed has a staining problem, so as to correct the staining problem of the image to be processed.

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 method according to any one of claims 1 to 7 is implemented.

10. An electronic device, characterized in that: The electronic device comprises: A memory storing a computer program; A processor is communicatively connected to the memory, and executes the method according to any one of claims 1 to 7 when calling the computer program.

Images