A Microscopic Image Enhancement and Restoration Method and System for Tumor Pathology

Abstract

The present invention discloses a microscopic image enhancement and restoration method and system for tumor pathology, which relates to the technical field of medical imaging. It includes: S1: Obtain the normalization value of the microscopic image, compare the normalization value with a preset normalization value, and determine the final microscopic image; S2: Determine the artifact region and artifact type corresponding to the final microscopic image, and process the artifact region; S3: Optimize the microscopic image after processing the artifact region, and reconstruct the three-dimensional vascular structure in the microscopic image to obtain the optimized microscopic image. The present invention uses an artifact detection model to identify artifacts in the preprocessed microscopic image, which can not only reduce the tiny noise existing in the microscopic image itself, but also improve the recognition efficiency and recognition result, so that the artifact region in the microscopic image can be better recognized, and further improve the restoration efficiency of the artifact region.

Description

Technical Field

[0001] The present invention relates to the technical field of medical imaging, and particularly to a microscopic image enhancement and restoration method and system for tumor pathology. Background Art

[0002] With the development of medical technology, the role of pathology in disease diagnosis and treatment has become increasingly prominent. Especially for complex diseases such as cancer, accurate pathological analysis is a prerequisite for formulating effective treatment plans. However, traditional pathology relies on doctors observing tissue sections through microscopes, which is not only time-consuming but also easily affected by subjective factors.

[0003] Currently, digital pathology has become an important research field. Digital microscopes can convert pathological images into electronic formats for easy storage, transmission, and analysis. Nevertheless, these images are often affected by various factors (such as noise, blurring), which will reduce the image quality and thus affect the judgment of pathologists. In addition, with the progress of high-resolution imaging technology, how to efficiently process a large amount of data has become a new challenge.

[0004] Chinese Patent Invention Publication No. CN112017131A discloses a method, device, and computer-readable storage medium for removing metal artifacts from CT images, including steps: collecting images with and without metal artifacts, and processing to obtain corresponding projection data; respectively reconstructing the generated two sets of projection data to generate images with and without metal artifacts, and further processing to obtain an initial artifact suppression image and a data integrity map; designing a neural network and a loss function, and the neural network outputs an image with artifacts removed, etc. The method for removing metal artifacts from CT images of this invention utilizes the correlation between metal artifacts and surrounding structures, as well as the relationship with tissue components, and overcomes the problem that existing image domain technologies cannot integrate the data integrity information in metal scan data into the processing process of the image domain, improving the accuracy of artifact correction and the quality of CT images.

[0005] During the process of processing pathological images with the above-mentioned and similar artifact removal methods, since pathological images usually contain complex noise components such as random noise, speckle noise, etc., these noises will seriously affect the image quality, resulting in the loss of key details and thus affecting the accuracy of diagnosis. At the same time, due to physical limitations in the sample preparation process or limitations in device performance, the acquired images often have varying degrees of blurring, further increasing the difficulty of pathological feature recognition. Summary of the Invention

[0006] The purpose of the present invention is to provide a microscopic image enhancement and restoration method and system for tumor pathology to solve the problems raised in the above background art.

[0007] To achieve the above object, the present invention provides the following technical solutions: A microscopic image enhancement and restoration method for tumor pathology, including:

[0008] S1: Image preprocessing: Obtain the normalization value of the microscopic image, and at the same time compare the normalization value with a preset normalization value, and determine the final microscopic image according to the comparison result;

[0009] S2: Image artifact processing: Identify the final microscopic image through an artifact detection model, determine the artifact area and artifact type corresponding to the final microscopic image, and at the same time process the artifact area according to the artifact type, including:

[0010] S2.1: Determine the artifact sorting: Identify the artifact area through an artifact detection model, determine the type and size corresponding to each artifact area, and sort all the artifact areas according to the type and size, including:

[0011] M1: Determine the artifact probability corresponding to each pixel through an artifact detection model, and at the same time convert the artifact probability into a binary image, and process the artifact area through the binary image and connected component labeling;

[0012] M2: Determine the type of each artifact area according to the characteristics of the artifact area;

[0013] M3: Determine the priority of the artifact area according to the area and artifact type of the artifact area;

[0014] S2.2: Perform artifact repair: Repair, remove and label the artifact area according to the sorting of the artifact area;

[0015] S2.3: Perform artifact detection: Detect the microscopic image after artifact processing to determine whether all the artifact areas in the microscopic image are repaired. When all the artifact areas are repaired, the next step is executed. Otherwise, repeat steps M1 - S2.3 until all the artifact areas are repaired;

[0016] S3: Optimize the microscopic image: Optimize the microscopic image after artifact area processing through multi-scale analysis, and reconstruct the three-dimensional vascular structure in the microscopic image through phase variance to obtain the optimized microscopic image.

[0017] Furthermore, determining the final microscopic image includes:

[0018] S1.1: Image processing: Smooth the microscopic image through a Gaussian filter, process the frequency of the microscopic image through histogram equalization, and normalize each pixel value of the microscopic image through linear normalization to obtain the corresponding normalized value of the microscopic image;

[0019] S1.2: Set the preset normalized value: Process the microscopic image through a CNN model to obtain the preset normalized value. Specifically:

[0020] ;

[0021] Where: is the preset normalized value, is the original pixel value of the i-th microscopic image at the position (x, y), is the coordinate of the pixel in the horizontal direction of the image, is the coordinate of the pixel in the vertical direction of the image, is the convolutional neural network, are all the trainable parameters of the convolutional neural network;

[0022] S1.3: Determine the processed image: Compare the obtained normalized value with the preset normalized value, obtain the difference between the normalized value and the preset normalized value, and compare the difference with the preset difference. When the difference is not greater than the preset difference, the microscopic image corresponding to the difference is the processed image; otherwise, repeat steps S1.1 and S1.3 until the obtained difference is not greater than the preset difference.

[0023] Furthermore, the acquisition formula for all the trainable parameters of the convolutional neural network is specifically:

[0024] ;

[0025] Where: is the value of all the trainable parameters at the (t + 1)-th iteration, is the value of all the trainable parameters at the t-th iteration, is the learning rate, is the loss function, is the gradient of the loss function with respect to all the trainable parameters.

[0026] Furthermore, sorting all the artifact regions includes:

[0027] M1: Determine the artifact regions: Through an artifact detection model, determine the artifact probability corresponding to each pixel, and at the same time convert the artifact probability into a binary image. Process the artifact regions through the binary image and connected component labeling;

[0028] M2: Determine the type of artifact: Based on the characteristics of the artifact region, determine the type of each artifact region, including circular or nearly circular artifacts, strip artifacts, and other artifacts;

[0029] M3: Determine the artifact ranking: Based on the area and type of the artifact region, determine the priority of the artifact region.

[0030] Furthermore, process the artifact region, including:

[0031] M1.1: Determine the initial artifact region: Through the artifact detection model, obtain the artifact probability corresponding to each pixel in the microscopic image, specifically:

[0032] ;

[0033] Where: is the probability that the pixel at position (x, y) belongs to an artifact, is the U-Net network structure, is the original pixel value of the i-th microscopic image at position (x, y), is the coordinate of the pixel in the horizontal direction of the image, is the coordinate of the pixel in the vertical direction of the image, are all trainable parameters of the U-Net model;

[0034] M1.2: Obtain the binary image: Convert the artifact probability corresponding to each pixel in the microscopic image into a binary image, and based on the binary image, determine the artifact region, specifically:

[0035] ;

[0036] Where: is the output binary image, is the probability that the pixel at position (x, y) belongs to an artifact, is the preset threshold;

[0037] M1.3: Determine the connected region: Perform connected component labeling on the artifact region, obtain the connected component label map and the area of each connected component, and compare the area of the connected component with the minimum area threshold to determine the final artifact region.

[0038] Furthermore, determine the final artifact region, including:

[0039] N1: Perform connected component labeling: Through the connected component labeling algorithm, label each artifact region to obtain the label map corresponding to the microscopic image;

[0040] N2: Obtain the area of the connected component: According to the number of pixels in each connected component, obtain the area of the connected component, specifically:

[0041] ;

[0042] Where: is the area of the i-th connected component, is the set of all pixel positions of the i-th connected component, is the coordinate of the pixel in the horizontal direction of the image, is the coordinate of the pixel in the vertical direction of the image;

[0043] N3: Remove small regions: Compare the area of the connected component with the minimum area threshold. When the area of the connected component is less than the minimum area threshold, delete the area of the connected component; otherwise, retain the area of the connected component.

[0044] N4: Merge adjacent artifact regions: Compare the minimum distance between two adjacent connected component regions with the preset distance. When the minimum distance is less than the preset distance, merge the two adjacent connected component regions corresponding to the minimum distance; otherwise, do not merge the two adjacent connected component regions. The formula for obtaining the minimum distance between two adjacent connected component regions is specifically:

[0045] ;

[0046] Where: is the connected component and the connected component is the minimum distance between them, is the set of all pixel positions of the i-th connected component, is the set of all pixel positions of the j-th connected component, is the coordinate of the pixel in the horizontal direction of the image within the i-th connected component, is the coordinate of the pixel in the horizontal direction of the image within the j-th connected component, is the coordinate of the pixel in the vertical direction of the image within the i-th connected component, is the coordinate of the pixel in the vertical direction of the image within the j-th connected component.

[0047] Furthermore, the determination method of the minimum area threshold is specifically:

[0048] According to the area of the connected component, obtain the histogram of the area of the connected component, and determine two main peaks from the histogram of the area of the connected component. The valley value between the two main peaks is the minimum area threshold.

[0049] Furthermore, obtaining the optimized microscopic image includes:

[0050] S3.1: Obtain the fused microscopic image: Smooth the microscopic image after processing the artifact area through Gaussian smoothing, downsample the smoothed microscopic image through downsampling to obtain a reduced microscopic image, and at the same time fuse the microscopic images at different scales to obtain a fused microscopic image;

[0051] S3.2: Reconstruct the three-dimensional vascular structure: Obtain the B-scan data of the fused microscopic image, determine the phase change at adjacent time points according to each B-scan position, and at the same time form a three-dimensional vascular structure according to the phase change.

[0052] A microscopic image enhancement and restoration system for tumor pathology uses a microscopic image enhancement and restoration method for tumor pathology described in any one of the above.

[0053] Compared with the prior art, the beneficial effects of the present invention are:

[0054] First: The present invention uses an artifact detection model to identify artifacts in the preprocessed microscopic image, which can not only reduce the tiny noise existing in the microscopic image itself, but also improve the recognition efficiency and recognition result, so that the artifact area in the microscopic image can be better recognized. At the same time, the type and size of the artifact area are recognized, and according to the recognition result, the artifact area is sorted, and according to the sorting result, corresponding artifact repair, removal and marking are carried out, thereby further improving the restoration efficiency of the artifact area;

[0055] Second: After the present invention preliminarily determines the artifact probability through the artifact detection model, the artifact probability is converted into a binary image, and the corresponding artifact area is determined according to the binary image. At the same time, through connected component labeling, the identified artifact area is further processed, thereby improving the accuracy of artifact area recognition. Description of the Drawings

[0056] Figure 1 It is a flow chart of the microscopic image enhancement and restoration method in the present invention;

[0057] Figure 2 It is a histogram of connected component areas in the present invention;

[0058] Figure 3 It is a flow chart of the process for determining the artifact area in the present invention. Detailed Embodiments

[0059] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0060] During the process of processing pathological images by existing artifact removal methods, since pathological images usually contain complex noise components, such as random noise, speckle noise, etc., these noises will seriously affect the image quality, resulting in the loss of key details, thereby affecting the accuracy of diagnosis. At the same time, due to physical limitations in the sample preparation process or limitations in equipment performance, the acquired images often have varying degrees of blurring, further increasing the difficulty of pathological feature recognition. The technical solution of this application identifies the preprocessed microscopic image through the constructed artifact detection model, determines the artifact area therefrom, and after sorting, repairing, and detecting the identified artifact area, detects the restored artifact area. At the same time, when the detection result shows that all artifacts have been removed, through multi-scale analysis, the microscopic image after the artifact area is restored is optimized, and the three-dimensional vascular structure is reconstructed through phase variance to obtain the processed microscopic image. That is to say, it can not only reduce the tiny noises existing in the microscopic image itself, but also improve the recognition efficiency and recognition result, so that the artifact area in the microscopic image can be better recognized.

[0061] Reference Figures 1-3 , this embodiment provides a microscopic image enhancement and restoration method for tumor pathology. The microscopic image enhancement and restoration method includes the following steps:

[0062] Step S1: Image preprocessing. That is, preprocess the microscopic image. That is to say, after performing image denoising, image enhancement, and image normalization processing, obtain the normalized value corresponding to the microscopic image, and at the same time compare the obtained normalized value with the preset normalized value, and determine the finally processed microscopic image according to the comparison result. Specifically as follows:

[0063] Step S1.1: Image processing. That is, smooth the microscopic image through a Gaussian filter so that the central pixel value of the microscopic image remains at the same value. Then, through histogram equalization, make the frequency of each gray level of the microscopic image remain within the preset range. At the same time, normalize each pixel value of the image through linear normalization. It should be noted that the image denoising, image enhancement, and image normalization processing in this embodiment are all existing conventional technical means, so they will not be specifically described in this embodiment.

[0064] Step S1.2: Set the preset normalization value. That is, process the microscopic image through the CNN model to obtain the corresponding preset normalization value of the microscopic image. Specifically:

[0065] ;

[0066] Where: is the preset normalization value, is the original pixel value of the i-th microscopic image at the position (x, y), is the coordinate of the pixel in the horizontal direction of the image, is the coordinate of the pixel in the vertical direction of the image, is the convolutional neural network, are all the trainable parameters of the convolutional neural network.

[0067] Furthermore, the formula for obtaining all the trainable parameters of the convolutional neural network is specifically:

[0068] ;

[0069] Where: is the value of all the trainable parameters at the (t + 1)-th iteration, is the value of all the trainable parameters at the t-th iteration, is the learning rate, is the loss function, is the gradient of the loss function with respect to all the trainable parameters.

[0070] Step S1.3: Determine the processed image. That is, take the microscopic image as the input of the CNN model in Step S1.2, output and obtain its corresponding preset normalization value. At the same time, perform image processing on the microscopic image according to Step S1.1 and obtain the corresponding image normalization value of the microscopic image. Furthermore, compare the image normalization value with the preset normalization value to obtain the difference size between the image normalization value and the preset normalization value. At the same time, compare the difference size with the preset difference. When the obtained difference is not greater than the preset difference, the obtained microscopic image is the final microscopic image. Otherwise, repeat Step S1.1 and Step S1.3 until the obtained difference is not greater than the preset difference. It should be noted that the preset difference can be set according to the actual needs in the actual use process, so it will not be specifically elaborated in this embodiment.

[0071] Step S2: Image artifact processing. That is, through the established artifact detection model, the preprocessed microscopic image in step S1.3 is identified, and the artifact regions in the microscopic image are determined. At the same time, the artifact regions are identified to determine the type corresponding to each artifact region, and the artifact regions are processed according to their corresponding artifact types. Specifically as follows:

[0072] Step S2.1: Determine the artifact sorting. That is, through the established artifact detection model, the artifact regions in the microscopic image are identified, the type and size corresponding to each artifact region are determined, and all the artifact regions are sorted according to the determined type and size. Specifically as follows:

[0073] Step M1: Determine the artifact regions. That is, through the constructed artifact detection model, the artifact probability corresponding to each pixel is determined. At the same time, each artifact probability is converted into a binary image, and the artifact regions are processed through the binary image and connected component labeling. Specifically as follows:

[0074] Step M1.1: Determine the artifact probability. That is, through the artifact detection model, the probability that each pixel in the microscopic image belongs to an artifact is obtained. Specifically:

[0075] ;

[0076] Where: is the probability that the pixel at position (x, y) belongs to an artifact, is the U-Net network structure, is the original pixel value of the i-th microscopic image at position (x, y), is the coordinate of the pixel in the horizontal direction of the image, is the coordinate of the pixel in the vertical direction of the image, are all the trainable parameters of the U-Net model.

[0077] That is to say, according to the artifact detection model, the probability that each pixel in the microscopic image belongs to an artifact can be obtained, and according to the obtained probability, the probability of the artifact region corresponding to the region where each pixel is located is determined.

[0078] In the process of specific implementation, a grayscale image with a size of 256*256 is used as the input of the artifact detection model, and its pixel range is [0,1]. Further, after being processed by the artifact detection model, it can output a probability map with a size of 256*256, where each pixel value represents the probability that the corresponding position belongs to an artifact. Specifically, the pixel value at position (100, 150) is 0.85, that is to say, the probability that the position (100, 150) belongs to the artifact region is 85%.

[0079] Step M1.2: Obtain a binary image. That is, convert the probability map obtained in Step M1.1 into a binary image, specifically as follows:

[0080] ;

[0081] Where: is the output binary image, is the probability that the pixel at position (x, y) belongs to an artifact, is a preset threshold value.

[0082] That is to say, obtain a binary image corresponding to the probability that each position belongs to the artifact area, and determine the artifact area according to the size of the binary image. Further, when the output binary image is 0, it indicates that the pixel does not belong to the artifact area; otherwise, it indicates that the pixel belongs to the artifact area.

[0083] In the process of specific implementation, for a 5*5 size probability map, some of its pixel values and the corresponding binary map values are shown in Table 1 below, specifically Table 1.

[0084] Table 1: Table of partial pixel values and binary map values corresponding to the probability map

[0085] ;

[0086] Further, the preset threshold value in this embodiment is set to 0.5. Therefore, pixels greater than 0.5 are marked as 1, that is, the area where the pixels marked as 1 are located is the artifact area, and pixels not greater than 0.5 are marked as 0, that is, the area where the pixels marked as 0 are located is not the artifact area.

[0087] Step M1.3: Determine the connected regions. That is, according to the artifact areas determined in Step M1.2, perform connected component labeling on each artifact area to obtain a connected component label map. At the same time, according to the connected component label map, obtain the area of each connected component, and determine the final artifact area based on the comparison between the area of the connected component and the minimum area threshold. Specifically as follows:

[0088] Step N1: Perform connected component labeling. That is, through a connected component labeling algorithm (such as breadth-first search, depth-first search, and union-find), set a label for each artifact area determined in Step M1.2, so that each artifact area corresponds to a connected component number, that is, obtain the corresponding label map.

[0089] Step N2: Obtain the area of the connected component. That is, obtain the number of pixels in each connected component and use the obtained number of pixels as the area within the corresponding connected component, specifically as follows:

[0090] ;

[0091] Wherein: is the area of the i-th connected region, is the set of all pixel positions of the i-th connected region, is the coordinate of the pixel in the horizontal direction of the image, is the coordinate of the pixel in the vertical direction of the image.

[0092] That is to say, by traversing the label map obtained in step N1, the number of occurrences of each label is determined, and at the same time, the number of pixels corresponding to each label is recorded through a hash table or an array.

[0093] Step N3: Remove small regions. That is, compare the area of the connected region obtained in step N2 with the minimum area threshold. When the area of the connected region is less than the minimum area threshold, the area of the connected region is deleted; otherwise, the area of the connected region is retained.

[0094] In this embodiment, according to the areas of all connected regions obtained in step N2, a corresponding connected region area histogram is obtained. At the same time, according to the connected region area histogram, two main peaks are determined, and the valley value between the two main peaks is used as the minimum area threshold.

[0095] In the process of specific implementation, a connected region area data set is set, specifically: [3, 5, 5, 7, 8, 8, 8, 10, 12, 14, 15, 15, 15, 18, 20, 22, 25, 30, 30, 30, 35, 40, 45, 50, 60]. At the same time, according to this connected region area data set, its corresponding connected region area histogram is obtained, as Figure 2 shown. Referring to Figure 2 it can be known that: the aggregation areas with smaller areas are: 5, 8, and 15, and the aggregation area with a larger area is: 30. That is to say, the two main peak regions are: the aggregation area with a smaller area (5, 8, 15) and the aggregation area with a larger area (30). Further, in this embodiment, the minimum area threshold is set to 25. That is, the areas of the connected regions to be deleted are: 3, 5, 5, 7, 8, 8, 8, 10, 12, 14, 15, 15, 15, 18, 20, and 22. The areas of the connected regions to be retained are: 25, 30, 30, 30, 35, 40, 45, 50, and 60.

[0096] Step N4: Merge adjacent artifact regions. That is, obtain the minimum distance between two adjacent connected component regions, and compare the obtained minimum distance with a preset distance (the preset distance can be specifically set according to actual requirements, so it is not specifically elaborated in this embodiment). When the minimum distance is less than the preset distance, merge the two adjacent connected component regions corresponding to the minimum distance. Otherwise, do not merge the two adjacent connected component regions.

[0097] In this embodiment, the minimum distance between two adjacent connected component regions is specifically:

[0098] ;

[0099] Where: is the connected component and the connected component the minimum distance between them, is the set of all pixel positions of the i-th connected component, is the set of all pixel positions of the j-th connected component, is the horizontal coordinate of the pixel in the i-th connected component in the image, is the horizontal coordinate of the pixel in the j-th connected component in the image, is the vertical coordinate of the pixel in the i-th connected component in the image, is the vertical coordinate of the pixel in the j-th connected component in the image.

[0100] In the process of specific implementation, the pixel positions included in the first connected component are: (1,1), (1,2), and (2,1). The pixel positions included in the second connected component are: (4,4), (4,5), and (5,4). Then the distances between each pair of corresponding pixels are: 4.24, 5, and 5. That is to say, the minimum distance between the first connected component and the second connected component is 4.24. Further, in the process of merging connected component regions, compare 4.24 with the preset distance.

[0101] Step M2: Determine the artifact type. That is, according to the characteristics (such as shape and texture) corresponding to the connected component regions determined in Step N4, determine the type of each artifact region, including circular or nearly circular artifacts, strip artifacts, and other artifacts.

[0102] Step M3: Determine the artifact sorting. That is, determine the priority of the artifact regions according to the area size and the corresponding artifact type of the artifact regions. Further, according to the artifact type determined in Step M2, sort the artifact regions in the order of circular or nearly circular artifacts, strip artifacts, and other artifacts. Then, according to all the artifact regions in each artifact type, sort them in descending order according to the artifact size.

[0103] Step S2.2: Perform artifact repair. That is, according to the artifact sorting determined in step M3, repair, remove, and mark the artifact regions. Specifically, through an interpolation method based on adjacent pixel information, repair the artifact regions to predict and fill in the missing pixel values. At the same time, set the artifact regions to the background color through binarization operations, or smooth the artifact regions through edge detection algorithms. Further, through marking, directly conduct manual review or perform specific processing according to the actual situation.

[0104] Step S2.3: Perform artifact detection. That is, detect the microscopic images processed in step S2.2 to determine whether all the artifact regions in each microscopic image have been repaired. When it is detected that the artifact regions in the microscopic images have not been repaired, repeat steps M1 - S2.3 until all the artifact regions have been repaired. Otherwise, execute the next step.

[0105] Step S3: Optimize the microscopic images. That is, through multi-scale analysis, optimize the microscopic images repaired in step S2.3, and reconstruct the three-dimensional vascular structure in the microscopic images through phase variance to obtain the optimized microscopic images. Specifically as follows:

[0106] Step S3.1: Obtain the fused microscopic images. That is, smooth the microscopic images repaired in step S2.3 through Gaussian smoothing, and then downsample the smoothed microscopic images through downsampling to obtain the reduced microscopic images. Further, fuse the microscopic images at different scales to obtain the fused microscopic images. It should be noted that the Gaussian smoothing process, downsampling process, and fusion process in this embodiment are all existing conventional technical means, so in this embodiment, no specific elaboration will be made.

[0107] Step S3.2: Reconstruct the three-dimensional vascular structure. That is, obtain continuous B-scan data through phase variance optical coherence tomography angiography. At the same time, according to each B-scan position, determine the phase change at adjacent time points and identify the blood flow signal from it. Further, combine all the phase variance maps to form a complete three-dimensional vascular structure.

[0108] This embodiment also provides a microscopic image enhancement and restoration system for tumor pathology, and this microscopic image enhancement and restoration system uses a microscopic image enhancement and restoration method for tumor pathology described in the above embodiment.

[0109] Although embodiments of the present invention have been shown and described, it will be understood by those of ordinary skill in the art that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the present invention, and the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for microscopic image enhancement and restoration for tumor pathology, characterized in that: Included are: S1: Image preprocessing: obtaining a normalized value of a microscopic image, comparing the normalized value with a preset normalized value, and determining a final microscopic image according to the comparison result; S2: Image artifact processing: The final microscopic image is identified by an artifact detection model, and the artifact area and artifact type corresponding to the final microscopic image are determined. At the same time, the artifact area is processed according to the artifact type, including: S2.1: Determine artifact sorting: Identify the artifact regions through an artifact detection model, determine the type and size corresponding to each artifact region, and sort all the artifact regions according to the type and size, including: M1: Determine the artifact probability corresponding to each pixel through the artifact detection model, and convert the artifact probability into a binary image. Process the artifact area through the binary image and the connected domain marker; M2: determining the type of each artifact region according to the characteristics of the artifact region; M3: determining the priority of the artifact region according to the area and the artifact type of the artifact region; S2.2: performing artifact repair: repairing, removing and marking the artifact areas according to the sorting of the artifact areas; S2.3: Performing artifact detection: Detecting the microscopic image after artifact processing to determine whether all artifact areas in the microscopic image are repaired. When all artifact areas are repaired, the next step is executed. Otherwise, steps M1 to S2.3 are repeated until all artifact areas are repaired. S3: Optimizing the microscopic image: Optimizing the microscopic image after the artifact area processing through multi-scale analysis, and reconstructing the three-dimensional vascular structure in the microscopic image through phase variance to obtain an optimized microscopic image.

2. A method for microscopic image enhancement and restoration for tumor pathology according to claim 1, characterized in that: Determine the final microscopic image, including: S1.1: Image processing: smoothing the microscopic image by using a Gaussian filter, processing the frequency of the microscopic image by using a histogram equalization, normalizing each pixel value of the microscopic image by using a linear normalization, and obtaining a normalized value corresponding to the microscopic image; S1.2: Setting a preset normalization value: Processing the microscopic image through the CNN model to obtain the preset normalization value, specifically: ； in: is the preset normalized value, is the original pixel value of the i-th microscopic image at position (x, y), is the horizontal coordinate of the pixel in the image, is the vertical coordinate of the pixel in the image, is a convolutional neural network, are all trainable parameters of the convolutional neural network; S1.3: Determine the processed image: compare the normalized value obtained with the preset normalized value, obtain the difference between the normalized value and the preset normalized value, and compare the difference with the preset difference. When the difference is not greater than the preset difference, the microscopic image corresponding to the difference is the processed image. Otherwise, repeat steps S1.1 and S1.3 until the difference obtained is not greater than the preset difference.

3. A method for microscopic image enhancement and restoration for tumor pathology according to claim 2, characterized in that: The formula for obtaining all trainable parameters of the convolutional neural network is as follows: ； in: are the values ​​of all trainable parameters at the t+1th iteration, are the values ​​of all trainable parameters at the tth iteration, is the learning rate, is the loss function, is the gradient of the loss function with respect to all trainable parameters.

4. A method for microscopic image enhancement and restoration for tumor pathology according to claim 3, characterized in that: Processing of artifact areas includes: M1.1: Determine the initial artifact area: Obtain the artifact probability corresponding to each pixel in the microscopic image through the artifact detection model, specifically: ； in: is the probability that the pixel at position (x, y) belongs to an artifact, is the U-Net network structure, is the original pixel value of the i-th microscopic image at position (x, y), is the horizontal coordinate of the pixel in the image, is the vertical coordinate of the pixel in the image, are all trainable parameters of the U-Net model; M1.2: Obtaining a binary image: Converting the artifact probability corresponding to each pixel in the microscopic image into a binary image, and determining the artifact area based on the binary image, specifically: ； in: is the output binary image, is the probability that the pixel at position (x, y) belongs to an artifact, is the preset threshold; M1.3: Determine the connected area: mark the connected domain of the artifact area, obtain the connected domain label map and the area of ​​each connected domain, and compare the connected domain area with the minimum area threshold to determine the final artifact area.

5. A method for microscopic image enhancement and restoration for tumor pathology according to claim 4, characterized in that: Determine the final artifact area, including: N1: Connected domain labeling: label each artifact region using a connected domain labeling algorithm to obtain a label map corresponding to the microscopic image; N2: Get the area of ​​the connected domain: Get the area of ​​the connected domain according to the number of pixels in each connected domain, specifically: ； in: is the area of ​​the i-th connected domain, is the set of all pixel positions of the i-th connected domain, is the horizontal coordinate of the pixel in the image, is the vertical coordinate of the pixel in the image; N3: Remove small areas: compare the area of ​​the connected domain with the minimum area threshold. When the area of ​​the connected domain is smaller than the minimum area threshold, the area of ​​the connected domain is deleted; otherwise, the area of ​​the connected domain is retained. N4: Merge adjacent artifact regions: Compare the minimum distance between two adjacent connected domain regions with the preset distance. When the minimum distance is less than the preset distance, merge the two adjacent connected domain regions corresponding to the minimum distance. Otherwise, do not merge the two adjacent connected domain regions. The formula for obtaining the minimum distance between two adjacent connected domain regions is as follows: ； in: Connected domain and connected domain The minimum distance between is the set of all pixel positions of the i-th connected domain, is the set of all pixel positions of the jth connected domain, is the horizontal coordinate of the pixel in the i-th connected domain, is the horizontal coordinate of the pixel in the jth connected domain, is the vertical coordinate of the pixel in the i-th connected domain, is the vertical coordinate of the pixel in the jth connected domain.

6. A method for microscopic image enhancement and restoration for tumor pathology according to claim 5, characterized in that: The method for determining the minimum area threshold is specifically as follows: According to the connected domain area, a connected domain area histogram is obtained, and two main peaks are determined from the connected domain area histogram, and the valley value between the two main peaks is the minimum area threshold.

7. The method for microscopic image enhancement and restoration for tumor pathology according to claim 1, characterized in that: Get optimized microscopy images, including: S3.1: Obtaining a fused microscopic image: smoothing the microscopic image after the artifact area processing by Gaussian smoothing, downsampling the smoothed microscopic image by downsampling to obtain a reduced microscopic image, and fusing microscopic images at different scales to obtain a fused microscopic image; S3.2: Reconstructing a three-dimensional vascular structure: Obtaining B-scan data of the fused microscopic image, and determining the phase change of adjacent time points according to each B-scan position, and constructing a three-dimensional vascular structure according to the phase change.

8. A microscopic image enhancement and restoration system for tumor pathology, characterized in that: A microscopic image enhancement and restoration method for tumor pathology as described in any one of claims 1 to 7 is used.

Images