Pathological tissue section dyeing equipment and method

Abstract

The invention provides pathological tissue section dyeing equipment and method. The method comprises the following steps: dividing a to-be-dyed lymphatic tissue section into a plurality of dyeing partitions according to the structural characteristics of lymphocyte tissues in the to-be-dyed lymphatic tissue section; dyeing agents with different concentrations are applied to different dyeing partitions according to the reference light transmittance of the lymphocyte tissue to obtain color gradient vectors of the lymphocyte tissue slices, light absorption characteristics of the dyeing partitions are measured through a photometer, and dyeing characteristics are determined based on the light absorption characteristics and the color gradient vectors; obtaining a dyeing image after dyeing the lymph tissue section, extracting deep features of the dyeing partitions from the dyeing image, and determining a dyeing loss value when dyeing the dyeing partitions based on the deep features and the dyeing features; and based on the dyeing loss value, carrying out self-adaptive adjustment on the dyeing agent concentration of each dyeing zone. By adopting the scheme provided by the invention, optimal adjustment on the concentration of the coloring agent in the sample dyeing process can be realized based on the characteristic difference of the cell tissues at different parts.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of sample staining, more particularly, the present application relates to a pathological tissue section staining device and method. BACKGROUND

[0002] Sample staining is a key technology in biomedical, pathological analysis and histological research. With the continuous deepening of medical research, the requirements for sample staining are also increasing. In the past, sample staining technology was relatively basic and could only present simple cell morphology. However, with the rapid development of science and technology, sample staining technology has made significant progress and can more clearly display cell structure and composition, providing strong support for accurate disease diagnosis and greatly promoting the improvement of medical diagnosis level.

[0003] In the prior art, sample staining is mainly achieved through traditional staining methods. For example, the conventional hematoxylin-eosin (HE) staining method is widely used in pathological diagnosis. This method can make the cell nucleus blue and the cytoplasm red, thereby distinguishing different cell structures. There is also an immunohistochemical staining technology that uses the specific binding of antigens and antibodies to stain and label specific proteins and other biological molecules. However, in terms of staining precision control, due to the complex and diverse structure of cell tissues, traditional staining methods are difficult to accurately stain different regions of cell tissues according to their characteristics. Single dye concentration leads to uneven staining results, and some subtle structures and pathological sites are difficult to clearly present, thereby affecting the accuracy of diagnosis and failing to meet the growing demand for precision medicine. Therefore, how to optimize and adjust the concentration of the dye during the sample staining process based on the differences in cell tissue characteristics of different parts has become a difficult problem in the industry. SUMMARY

[0004] The present application provides a pathological tissue section staining device and method, which can optimize and adjust the concentration of the dye during the sample staining process based on the differences in cell tissue characteristics of different parts.

[0005] In a first aspect, the present application provides a pathological tissue section staining control method, comprising the following steps: Obtaining a lymph tissue section to be stained; According to the structural characteristics of the lymphocyte tissue in the lymph tissue section, the lymph tissue section is divided into a plurality of staining sub-zones; According to the reference light transmittance of the lymphocyte tissue in each staining sub-zone, different concentrations of dye are applied to each staining sub-zone for initial staining, and then a color gradient vector of the lymph tissue section is obtained. The light absorption characteristics of each staining sub-zone are determined by a photometer, and then the staining characteristics of the lymph tissue section are determined based on all the light absorption characteristics and the color gradient vector. obtaining a staining image of the lymph tissue slice after staining, extracting deep features of each staining sub-region from the staining image, and determining a staining loss value of each staining sub-region based on the deep features and the staining features; performing adaptive adjustment on the concentration of the staining agent of each staining sub-region based on the staining loss value.

[0006] In some embodiments, dividing the lymph tissue slice into a plurality of staining sub-regions according to structural features of lymphocyte tissues in the lymph tissue slice specifically includes: obtaining a whole field image of the lymph tissue slice; extracting structural features of lymphocyte tissues in the lymph tissue slice from the whole field image; dividing the lymph tissue slice into a plurality of staining sub-regions based on the structural features.

[0007] In some embodiments, performing initial staining on each staining sub-region by applying different concentrations of staining agents according to a reference light transmittance of lymphocyte tissues in each staining sub-region, and then obtaining a color gradient vector of the lymph tissue slice specifically includes: constructing a mapping relationship database between light transmittance and concentration of staining agents; determining a reference light transmittance of lymphocyte tissues in each staining sub-region; performing matching query in the mapping relationship database according to the reference light transmittance, and obtaining an adaptive concentration of staining agents of each staining sub-region; performing staining on cell tissues in each staining sub-region based on the adaptive concentration of staining agents corresponding to each staining sub-region, and then obtaining a stained slice; extracting a color gradient vector of the lymph tissue slice based on the stained slice.

[0008] In some embodiments, determining staining features of the lymph tissue slice based on all light absorption features and the color gradient vector specifically includes: constructing high-dimensional features according to all light absorption features and the color gradient vector; extracting staining features of the lymph tissue slice from the high-dimensional features.

[0009] In some embodiments, extracting deep features of each staining sub-region from the staining image specifically includes: performing sub-region positioning on all staining sub-regions in the staining image, and obtaining a plurality of sub-image blocks, each sub-image block corresponding to a staining sub-region; extracting features of each sub-image block respectively, and then obtaining deep features of each staining sub-region.

[0010] In some embodiments, the adaptive adjustment of the dyeing agent concentration of each dyeing sub-area based on the dyeing loss value specifically comprises: determining a control error corresponding to the dyeing loss value; obtaining an adaptive dyeing agent concentration corresponding to each dyeing sub-area; updating the dyeing agent concentration of each dyeing sub-area based on the adaptive dyeing agent concentration and the control error, so as to realize the adaptive adjustment of the dyeing agent concentration.

[0011] In some embodiments, the lymphoid tissue section is obtained by using a standardized pathological tissue section preparation process.

[0012] In a second aspect, the present application provides a pathological tissue section dyeing device, comprising a dyeing control unit, wherein the dyeing control unit comprises: an obtaining module, configured to obtain a lymphoid tissue section to be dyed; a processing module, configured to divide the lymphoid tissue section into a plurality of dyeing sub-areas according to the structural features of lymphocyte tissues in the lymphoid tissue section; the processing module is further configured to apply different concentrations of dyeing agents to each dyeing sub-area for initial dyeing according to the reference light transmittance of lymphocyte tissues in each dyeing sub-area, so as to obtain a color gradient vector of the lymphoid tissue section, determine the light absorption characteristics of each dyeing sub-area by using a luminometer, and then determine the dyeing characteristics of the lymphoid tissue section based on all the light absorption characteristics and the color gradient vector; the processing module is further configured to obtain a dyeing image of the lymphoid tissue section after dyeing processing, extract deep features of each dyeing sub-area from the dyeing image, and determine a dyeing loss value of each dyeing sub-area based on the deep features and the dyeing characteristics; an executing module, configured to perform adaptive adjustment of the dyeing agent concentration of each dyeing sub-area based on the dyeing loss value.

[0013] In a third aspect, the present application provides a computer device, comprising a memory and a processor, wherein the memory stores a code, and the processor is configured to obtain the code and execute the pathological tissue section dyeing control method described above.

[0014] In a fourth aspect, the present application provides a computer readable storage medium, wherein the computer readable storage medium stores a computer program, and the computer program is executed by a processor to realize the pathological tissue section dyeing control method described above.

[0015] The technical scheme provided by the embodiments of the present application has the following beneficial effects: The pathological tissue section staining device and method provided by the application first obtains a lymph tissue section to be stained; secondly, the lymph tissue section is divided into multiple staining sub-zones according to the structural characteristics of lymphocyte tissues in the lymph tissue section; further, different concentrations of staining agents are respectively applied to each staining sub-zone for initial staining according to the reference light transmittance of lymphocyte tissues in each staining sub-zone, and then a color gradient vector of the lymph tissue section is obtained; the light absorption characteristics of each staining sub-zone are determined by a photometer, and then the staining characteristics of the lymph tissue section are determined based on all the light absorption characteristics and the color gradient vector; then, a staining image after staining processing of the lymph tissue section is obtained, the deep features of each staining sub-zone are extracted from the staining image, and the staining loss value of each staining sub-zone during staining is determined based on the deep features and the staining characteristics; finally, the concentration of the staining agent of each staining sub-zone is adaptively adjusted based on the staining loss value.

[0016] It can be seen that the application can realize the optimal adjustment of the concentration of the staining agent in the sample staining process based on the differences in the characteristics of the cell tissues in different parts. First, the lymph tissue section is divided into multiple staining sub-zones according to the structural characteristics of the lymphocyte tissues in the lymph tissue section to realize local staining analysis, thereby avoiding uneven staining effect caused by a single staining agent concentration. Second, the staining sub-zones are subjected to initial staining by applying staining agents with adaptive concentrations according to the reference light transmittance of the lymphocyte tissues in the staining sub-zones, and then a color gradient vector of the lymph tissue section is obtained to make the staining area present a better color development effect after staining, thereby better reflecting the color transition between different areas and being conducive to optimizing the staining parameters. Further, the staining characteristics of the lymph tissue section are determined based on the light absorption characteristics of the staining sub-zones and the color gradient vector to comprehensively express the optical and color characteristics in combination with the light absorption characteristics of the cell tissues in the staining sub-zones after staining processing, thereby being conducive to evaluating the staining quality of the section and optimizing the subsequent staining process. Then, the deep features of the staining sub-zones are extracted from the staining image after staining processing of the lymph tissue section to describe the fine structure, texture pattern and color distribution of the lymph tissue section, more essentially describe the characteristics of the staining sub-zones, thereby being conducive to subsequent analysis and understanding of the lymph tissue section. In addition, the staining loss value of the staining sub-zones during staining is determined based on the deep features and the staining characteristics to reflect the difference degree between the staining characteristics of the staining sub-zones and the overall staining characteristics of the lymph tissue section, thereby being conducive to optimizing the subsequent staining parameters. Finally, the concentration of the staining agent of each staining sub-zone is adaptively adjusted based on the staining loss value to make the staining effect in each staining sub-zone close to the expected target. In summary, the technical solution provided by the application can realize the optimal adjustment of the concentration of the staining agent in the sample staining process based on the differences in the characteristics of the cell tissues in different parts. BRIEF DESCRIPTION OF DRAWINGS

[0017] Figure 1 is an example flow chart of a pathological tissue section staining control method according to some embodiments of the present application; Figure 2 is an example flow chart of preparing a lymphatic tissue section according to some embodiments of the present application; Figure 3 is an example flow chart of determining light absorption characteristics according to some embodiments of the present application; Figure 4 is a structural schematic diagram of a staining control unit according to some embodiments of the present application; Figure 5 is a structural schematic diagram of a computer device implementing a pathological tissue section staining control method according to some embodiments of the present application. DETAILED DESCRIPTION

[0018] In order to better understand the technical solutions of the present application, the technical solutions of the present application will be described in detail below in combination with the drawings of the specification and specific embodiments.

[0019] Referring to Figure 1 , the figure is an example flow chart of a pathological tissue section staining control method according to some embodiments of the present application, which mainly includes the following steps: In step 101, a lymphatic tissue section to be stained is obtained.

[0020] In specific implementation, the lymphatic tissue section to be stained is obtained, that is, the lymphatic tissue section to be stained is obtained through a standardized pathological tissue section preparation process. Specifically, in this embodiment, the target lymphatic tissue is first cut off using a scalpel and immediately placed in 10% neutral buffered formalin for fixation. Then, the target lymphatic tissue sample is sequentially dehydrated by gradient alcohol, transparentized by xylene, and infiltrated and embedded in molten paraffin. Next, the target lymphatic tissue embedded in paraffin is precisely cut into a thin slice of 3-5µm using a rotary microtome to obtain a lymphatic tissue section, which is spread on a glass slide coated with a protein adhesive. Finally, the lymphatic tissue section is dewaxed by xylene and rehydrated by gradient ethanol to remove paraffin residues, so that the lymphatic tissue section is suitable for staining treatment, thereby obtaining the lymphatic tissue section to be stained, as shown in Figure 2 , the figure is an example flow chart of preparing a lymphatic tissue section according to some embodiments of the present application, and in addition, in other embodiments, other process combinations can also be used to obtain a lymphatic tissue section, which is not limited here.

[0021] In step 102, the lymphatic tissue section is divided into a plurality of staining sub-zones according to the structural characteristics of the lymphocyte tissue in the lymphatic tissue section.

[0022] In some embodiments, the lymph tissue slice is divided into multiple staining sub-zones according to structural features of lymphocyte tissue in the lymph tissue slice in the following manner: Obtaining a full-field image of the lymph tissue slice; Extracting structural features of lymphocyte tissue in the lymph tissue slice from the full-field image; Regionally dividing the lymph tissue slice based on the structural features to obtain multiple staining sub-zones.

[0023] In a specific implementation, the full-field image of the lymph tissue slice is obtained by an optical scanner, and in other embodiments, other devices such as a high-resolution microscope can also be used to obtain the full-field image, which is not limited here.

[0024] In a specific implementation, the structural features of lymphocyte tissue in the lymph tissue slice are extracted from the full-field image, that is, a 100*100 pixel window is taken as a processing unit, and the cell nucleus density and fibrosis area proportion of lymphocyte tissue in the processing unit are extracted from the full-field image by a deep learning-based segmentation model, and then multiple cell nucleus densities and fibrosis area proportions are obtained. All extracted cell nucleus densities and fibrosis area proportions are combined as the structural features of lymphocyte tissue in the lymph tissue slice. Specifically, in this embodiment, a U-Net model is first selected as the deep learning-based segmentation model. Then, each processing unit is processed by the model, and the model output is a pixel-level prediction of the cell nucleus and fibrosis area in each processing unit. Further, the pixel number of the cell nucleus area in each processing unit is counted by Open CV, and then divided by the pixel area of the corresponding processing unit to obtain the cell nucleus density of each processing unit. In addition, the pixel number of the fibrosis area in each processing unit is counted to obtain the fibrosis area proportion of each processing unit.

[0025] It should be noted that the structural features in this application represent the characteristic parameters of the microstructure of the cell tissue in the slice. Specifically, the structural features are represented by the cell nucleus density and the fibrosis area proportion in this application. By extracting the structural features, the subsequent staining process can be optimized.

[0026] In a specific implementation, the lymph tissue slice is regionally divided based on the structural features to obtain a plurality of staining sub-regions, that is, the cell nucleus density and the fibrosis area proportion corresponding to each processing unit are obtained from the structural features, all processing units are divided by a double threshold method, and a plurality of staining sub-regions are obtained. The specific threshold can be set according to actual application requirements combined with historical practical experience, which is not limited here. The staining sub-region refers to a local region divided according to different tissue structures. Each staining sub-region corresponds to different cell tissue components. Dividing the staining sub-region is beneficial to analyzing the local staining intensity, uniformity and biological characteristics to optimize the subsequent staining parameters.

[0027] In step 103, each staining sub-region is subjected to initial staining by applying different concentrations of staining agents according to the reference light transmittance of the lymphocyte tissue in each staining sub-region, and then a color gradient vector of the lymph tissue slice is obtained. The light absorption characteristics of each staining sub-region are determined by a photometer, and then the staining characteristics of the lymph tissue slice are determined based on all the light absorption characteristics and the color gradient vector.

[0028] In some embodiments, each staining sub-region is subjected to initial staining by applying different concentrations of staining agents according to the reference light transmittance of the lymphocyte tissue in each staining sub-region, and then a color gradient vector of the lymph tissue slice is obtained. The following methods can be used, that is: A mapping relationship database of light transmittance and staining agent concentration is constructed. The reference light transmittance of the lymphocyte tissue in each staining sub-region is determined. The reference light transmittance is matched and queried in the mapping relationship database to obtain the adaptive staining agent concentration of each staining sub-region. The cell tissues in each staining sub-region are stained based on the adaptive staining agent concentration corresponding to each staining sub-region, and then a stained slice is obtained. The color gradient vector of the lymph tissue slice is extracted based on the stained slice.

[0029] In a specific implementation, a mapping relationship database of light transmittance and dye concentration is constructed, that is, first, the light transmittance of a large number of unstained lymph tissue sections is measured using a high-resolution microscope, and the optical density distribution is obtained through full-field scanning imaging; then, based on the light transmittance of the unstained lymph tissue sections, the tissue regions are classified, and the unstained lymph tissue sections are divided into different light transmittance partitions using the K-means clustering method; secondly, under experimental conditions, the dye prepared by the standard concentration gradient method is applied to different light transmittance partitions, and the stained tissue images are analyzed by computer vision algorithm to quantify the staining uniformity, contrast and color saturation of each partition to evaluate the staining quality; then, based on the evaluation results of the staining quality, a least squares regression model is used to establish the mapping relationship between the light transmittance and the concentration of the dye, and the mapping relationship is stored in the database, and the metadata such as the type of the experimental sample and the staining conditions are recorded for subsequent optimization and expansion, and thus the mapping relationship database of light transmittance and dye concentration is obtained.

[0030] In a specific implementation, the reference light transmittance of lymphocyte tissue in each staining partition is determined, that is, first, a standard white light is used to irradiate the lymph tissue section to ensure uniformity of light intensity, and a full-field image of the lymph tissue section is obtained; secondly, an automatic exposure correction algorithm of histogram equalization is used to preprocess the full-field image to eliminate the error influence of the device; then, for each staining partition, the reference light transmittance is determined by using optical transmission measurement technology, specifically, the mean value of the gray value of each pixel in the staining partition is calculated as the transmitted light intensity, and the mean value of the gray value of the blank area (i.e. the sample-free area) is calculated as the reference light intensity, and the transmitted light intensity and the reference light intensity are used as input parameters to substitute into the optical density conversion formula, and thus the reference light transmittance of the lymphocyte tissue in each staining partition is obtained.

[0031] It should be noted that the reference light transmittance in the present application represents an index for evaluating the light transmittance of the tissue region, specifically, the reference light transmittance in the present application refers to the light transmittance of the unstained lymphocyte tissue region, by determining the reference light transmittance, the density, refractive index and internal structure light scattering characteristics of the cell tissue in the section can be reflected, which can provide a basis for subsequent dye concentration adjustment, thereby realizing accurate staining control.

[0032] In a specific implementation, the reference light transmittance is matched and queried in the mapping relationship database to obtain the adaptive dyeing agent concentration of each dyeing partition, that is, the reference light transmittance is input as a query parameter into the mapping relationship database for matching and searching, if the reference light transmittance exists in the mapping relationship database, the corresponding dyeing agent concentration is directly returned as the adaptive dyeing agent concentration, if the reference light transmittance is not explicitly stored in the mapping relationship database, an interpolation algorithm such as a linear interpolation algorithm can be used to calculate the adaptive dyeing agent concentration, and then the adaptive dyeing agent concentration of each dyeing partition is obtained.

[0033] It should be noted that the adaptive dyeing agent concentration in this embodiment represents the dyeing agent concentration matched for the cell tissue characteristics in a specific dyeing area. By determining the adaptive dyeing agent concentration, the dyeing area can exhibit a better color development effect after dyeing, thereby improving the contrast and resolution of the pathological tissue section for subsequent analysis and diagnosis.

[0034] In a specific implementation, the cell tissues in each dyeing partition are dyed based on the adaptive dyeing agent concentration corresponding to each dyeing partition, and then a dyeing section is obtained, that is, the dyeing operation is performed by a multi-channel micro-injection pump to obtain the dyeing section. Specifically, each channel of the multi-channel micro-injection pump corresponds to a dyeing partition, the dyeing agent solution with the adaptive dyeing agent concentration corresponding to each dyeing partition is loaded into the corresponding liquid storage tank, and the dyeing operation is completed through a computer control interface.

[0035] In a specific implementation, a color gradient vector of the lymph tissue section is extracted based on the dyeing section, that is, first, a high-resolution digital microscope is used to collect digital images of the dyeing section, and the collected images are converted from an RGB color space to a Lab color space by means of Open CV; second, the image segmentation algorithm is used to segment the image converted in the color space into multiple sub-regions according to the division method of the dyeing partition, one dyeing partition corresponding to one sub-region; then, for each dyeing partition, the Sobel gradient operator is used to calculate the gradient amplitudes of the color components L (brightness), a (red-green axis), and b (yellow-blue axis) in the corresponding sub-region, and the calculation results are combined to form a color feature vector; finally, the color feature vectors corresponding to all dyeing partitions are concatenated to obtain the color gradient vector of the lymph tissue section.

[0036] It should be noted that the color gradient vector in this application represents the change of the color of the image in space, specifically, the color gradient vector in this application refers to the color space distribution of the dyed lymph tissue section. By determining the color gradient vector, the color transition between different regions can be reflected, thereby facilitating the optimization of dyeing parameters.

[0037] In addition, it should be noted that the light absorption characteristics in this application represent the characteristics of the absorption capacity and absorption mode of a substance for light of different wavelengths. Specifically, the light absorption characteristics in this application reflect the light absorption characteristics of the cell tissue in each stained partition after staining. This characteristic is determined by factors such as the composition and structure of the cell tissue itself and the binding state of the dye and the cell tissue. Measuring the light absorption characteristics is beneficial for evaluating the staining quality of the slice, thereby optimizing the subsequent staining process. As a preferred embodiment, reference is made to Figure 3 As shown in FIG. 1 , this figure is an exemplary flow chart of determining light absorption characteristics according to some embodiments of the present application. In this embodiment, the light absorption characteristics of each stained partition can be measured by a photometer using the following steps: First, in step 1031, a spectrophotometer with a wavelength accuracy of ±0.5 nm, an absorbance measurement range of 0-3 Abs, and an accuracy of ±0.02 Abs is selected and preheated for no less than 30 minutes; Next, in step 1032, the wavelength is calibrated using a standard filter with an error controlled within ±0.1 nm, and the absorbance of the photometer is calibrated to 0 using a blank glass slide as a reference; Then, in step 1033, the stained section is placed on a sample stage and observed with a microscope. The X, Y, and Z axes of the sample stage are adjusted to position each stained section so that each stained section is located at the center of the measurement light path, and the focal length is adjusted to ensure a clear image within the microscope field of view. Then, in step 1034, a wavelength scanning range of 400-700 nm and a scanning interval of 3 nm are selected, the photometer is started, each stained partition is measured, the absorbance at each scanning wavelength is recorded, and the measurement is repeated 5 times to obtain an average value to reduce system error; Finally, in step 1035, outliers are removed according to the 3σ principle, the light absorption spectrum curve corresponding to each dyeing partition is drawn, and the maximum absorption wavelength and the maximum absorption peak absorbance value are extracted, and the extracted characteristic parameters are combined into the light absorption characteristics of each dyeing partition.

[0038] In some embodiments, the staining characteristics of the lymphatic tissue section may be determined based on all light absorption characteristics and the color gradient vector in the following manner, namely: Constructing a high-dimensional feature based on all light absorption features and the color gradient vector; The staining features of the lymphatic tissue section are extracted from the high-dimensional features.

[0039] In a specific implementation, a high-dimensional feature is constructed according to all light absorption features and the color gradient vector, that is, the high-dimensional feature is obtained by fusing all light absorption features and the color gradient vector. Specifically, first, the color feature vector corresponding to each staining subregion is extracted from the color gradient vector by using the Numpy component in Python, the extracted color feature vector is connected with the corresponding light absorption feature to obtain the fusion feature of each staining subregion; then, all fusion features are matrixed and aligned, and the matrix obtained after the alignment is taken as the high-dimensional feature, so as to comprehensively integrate information in the color space and the light absorption space, and each row in the high-dimensional feature represents the fusion feature of a staining subregion.

[0040] In a specific implementation, the staining feature of the lymph tissue section is extracted from the high-dimensional feature, that is, the high-dimensional feature is processed by dimension reduction to remove redundant information and extract key feature information, and the extracted feature information is taken as the staining feature of the lymph tissue section. Specifically, first, the Z-score standardization method is used to normalize each column of feature data in the high-dimensional feature, so as to avoid errors caused by inconsistent data scales on the dimension reduction result; then, the PCA method in sklearn.decomposition.PCA in Python is used to gradually extract principal components with maximum variance from the normalized high-dimensional feature, and the n_components parameter is set to specify the number of principal components to be retained. The number of principal components can be determined by using the variance explanation rate method, that is, by gradually increasing the number of principal components, the cumulative variance explanation rate is calculated, and when the cumulative variance explanation rate reaches a certain threshold, the corresponding number of principal components can meet the preset n_components value. For example, the threshold is set to PCA (n_components=0.95), that is, the data after dimension reduction can still retain 95% of the original information; further, the fit_transform method of the PCA object is called to process the normalized high-dimensional feature, which identifies the extracted principal components and converts the high-dimensional feature into a low-dimensional feature according to the extracted principal components; finally, the principal component contribution method (pca.components_) is used to select high-contribution features that can best represent the staining effect of the tissue from the low-dimensional feature, and the extracted high-contribution features are taken as the staining feature of the lymph tissue section.

[0041] It should be noted that the staining feature in the present application represents the comprehensive expression of optical and color characteristics of the lymph tissue section after staining, and the determination of the staining feature is beneficial to the implementation of the subsequent staining agent concentration optimization process.

[0042] In step 104, a staining image of the lymph tissue slice after staining is obtained, deep features of each staining sub-region are extracted from the staining image, and a staining loss value when each staining sub-region is stained is determined based on the deep features and the staining features.

[0043] In a specific implementation, the staining image of the lymph tissue slice after staining is obtained by placing the lymph tissue slice on a microscope stage, adjusting the magnification and focal length of the microscope, connecting the microscope with a scientific CMOS camera to collect a high-definition RGB image, and thus obtaining the staining image of the lymph tissue slice after staining. The staining image refers to a digital image of the tissue slice obtained under a microscope. The staining image can reflect the morphological structure, cell distribution, and staining characteristics of the cell tissue in the lymph tissue slice.

[0044] In some embodiments, the deep features of each staining sub-region are extracted from the staining image in the following manner: All staining sub-regions in the staining image are located and positioned, and a plurality of sub-image blocks are obtained, each corresponding to a staining sub-region. The deep features of each staining sub-region are obtained by extracting features from each sub-image block respectively.

[0045] In a specific implementation, all staining sub-regions in the staining image are located and positioned to obtain a plurality of sub-image blocks in the following manner: first, the staining image is preprocessed by color normalization to reduce the color difference between staining batches and applying Gaussian filtering to enhance image details; then, the staining image is segmented at the pixel level using a deep learning-based image segmentation method, such as DeepLabV3+, to obtain the region mask of each staining sub-region; further, the boundaries of each staining sub-region are accurately identified using a contour detection method, such as findContours in Open CV, and a corresponding polygon boundary box is generated to distinguish different staining sub-regions; finally, the staining image is cropped using the corresponding polygon boundary box to obtain a plurality of sub-image blocks for subsequent feature extraction and analysis.

[0046] In a specific implementation, the deep features of each staining sub-region are obtained by extracting features from each sub-image block respectively in the following manner: a deep convolutional neural network model is used to encode features of each sub-image block to obtain the deep features of each staining sub-region. Specifically, a pre-trained ResNet50 model is used to extract multi-level texture, color, and morphological features of the slice tissue in each sub-image block, and the features are fused through a numpy component. The high-dimensional features obtained by fusion are used as the deep features of each staining sub-region.

[0047] It should be noted that the deep feature representation in the present application describes the high-level representation information of the stained section tissue, and specifically, the deep feature in the present application refers to the high-level representation information describing the fine structure, texture pattern and color distribution of the lymph tissue section extracted from each staining sub-region of the staining image, and the deep feature is automatically learned through the multi-layer structure of the deep convolutional neural network, can more abstractly and essentially describe the characteristics of the staining sub-region, has higher accuracy and robustness, and is beneficial to subsequent analysis and understanding of the lymph tissue section.

[0048] In some embodiments, determining the staining loss value when staining each staining sub-region based on the deep feature and the staining feature can adopt the following manner, that is: determining a plurality of feature difference indicators between the deep feature and the staining feature; determining the staining loss value when staining each staining sub-region based on all the feature difference indicators.

[0049] In some embodiments, determining a plurality of feature difference indicators between the deep feature and the staining feature can adopt the following manner, that is: aligning the feature dimensions of the deep feature and the staining feature; determining a plurality of feature difference indicators between the deep feature and the staining feature after the feature dimension alignment.

[0050] In specific implementation, the feature dimension alignment of the deep feature and the staining feature is performed through the feature mapping method based on the Lagrange interpolation method, and specifically, first, the Min-Max method is used to normalize the deep feature and the staining feature respectively to ensure the consistency of the numerical range; then, the interpolation point set is constructed in different dimensional spaces, and the Lagrange interpolation polynomial is established for the deep feature and the staining feature respectively, the interpolation in the unified dimension is calculated, and thus the aligned feature representation is obtained.

[0051] It should be noted that only the Lagrange interpolation method is called in the present embodiment, and the detailed implementation process will not be described here, and in addition, in other embodiments, other methods can also be used for feature alignment, for example, nearest neighbor interpolation, etc., which is not limited here. The feature dimension alignment can optimize the subsequent calculation process, thereby improving the accuracy of analysis.

[0052] In a specific implementation, the feature difference indicators between the deep features and the staining features after the feature dimension alignment are determined, that is, the feature distance values between the deep features and the staining features after the feature dimension alignment are calculated according to the Euclidean distance formula, the calculated feature distance is taken as the feature difference indicator of each feature dimension, and then the feature difference indicators between the deep features and the staining features are obtained, which describe the feature difference degree of the original tissue characteristics of the lymph tissue section and the tissue characteristics after staining in the corresponding dimension.

[0053] In a specific implementation, the staining loss value when each staining subregion is stained is determined based on all the feature difference indicators, that is, the weighted distance calculation is performed on all the feature difference indicators, and the calculation result is taken as the staining loss value when each staining subregion is stained. Specifically, in this embodiment, the weight of each feature dimension is determined based on the feature importance score algorithm of the random forest, a higher weight is given to the feature dimension that plays a key role in distinguishing different staining effects, and a lower weight is given to the feature dimension with lower correlation, and the sum of the weights of all feature dimensions is 1. In addition, in other embodiments, other methods can be used to assign weights, for example, expert experience method, feature selection method in machine learning, etc., which are not limited here.

[0054] It should be noted that the staining loss value in this application represents a quantitative indicator for evaluating the staining quality of each staining subregion, the smaller the staining loss value, the higher the staining quality of the corresponding staining subregion, and the staining effect of the staining subregion can better represent the staining effect of the whole lymph tissue section, the larger the staining loss value, the lower the staining quality of the corresponding staining subregion. By determining the staining loss value, the difference degree between the staining feature of the staining subregion and the whole staining feature of the lymph tissue section can be reflected, thereby facilitating the optimization of subsequent staining parameters.

[0055] In step 105, the dye concentration of each staining subregion is adaptively adjusted based on the staining loss value.

[0056] In some embodiments, the dye concentration of each staining subregion is adaptively adjusted based on the staining loss value in the following manner, that is: determine the control error corresponding to the staining loss value; obtain the adaptive dye concentration corresponding to each staining subregion; update the dye concentration of each staining subregion based on the adaptive dye concentration and the control error to realize adaptive adjustment of the dye concentration.

[0057] In a specific implementation, the control error corresponding to the dyeing loss value is determined, that is, an error threshold is set, and a difference between the error threshold and the dyeing loss value is taken as the control error corresponding to the dyeing loss value. The control error represents a deviation between a current output value of the system and an expected target value. Specifically, in this embodiment, the control error refers to a difference between a target dyeing loss value and a current dyeing loss value. When the control error is positive, it indicates that the current dyeing concentration is insufficient, and the dyeing agent concentration needs to be increased. When the control error is negative, it indicates that the dyeing is excessive, and the dyeing agent concentration needs to be reduced. The dyeing agent concentration is dynamically adjusted through the control error, so that the control error meets the system requirements, thereby achieving dyeing optimization.

[0058] In a specific implementation, the dyeing agent concentration of each dyeing sub-area is updated based on the adaptive dyeing agent concentration and the control error, so as to achieve adaptive adjustment of the dyeing agent concentration. That is, if the control error is positive, the dyeing concentration of the corresponding dyeing sub-area is insufficient, and the adaptive dyeing agent concentration is increased by a magnitude proportional to the size of the control error. Specifically, the increased dyeing agent concentration is the control error multiplied by a pre-set proportional coefficient. The size of the proportional coefficient can be determined according to historical tissue slice dyeing experiments, which is not limited here, so as to obtain the updated dyeing agent concentration of the dyeing sub-area. If the control error is negative, the dyeing concentration of the corresponding dyeing sub-area is too high, and the adaptive dyeing agent concentration is reduced by a magnitude proportional to the size of the absolute value of the control error. The reduced dyeing agent concentration value is also the absolute value of the control error multiplied by the proportional coefficient, so as to update the dyeing agent concentration of the dyeing sub-area. Through the above dynamic adjustment mode, the control error is continuously made to meet the system requirements for the dyeing loss value, thereby achieving adaptive adjustment of the dyeing agent concentration of each dyeing sub-area, and further making the dyeing effect in each dyeing sub-area close to the expected target.

[0059] In addition, another aspect of the present application provides a pathological tissue slice dyeing device, which includes a dyeing control unit, and the dyeing control unit is configured to perform the method for controlling the dyeing of a lymphoid tissue slice according to the above embodiments. Figure 4 The figure is a structural schematic diagram of a dyeing control unit according to some embodiments of the present application. The dyeing control unit 200 includes an acquisition module 201, a processing module 202, and an execution module 203, which are described as follows. The acquisition module 201 is mainly used for acquiring a lymphoid tissue slice to be dyed in the present application. The processing module 202 is mainly used for dividing the lymphoid tissue slice into a plurality of dyeing sub-areas according to the structural characteristics of lymphocyte tissues in the lymphoid tissue slice in the present application. The processing module 202 is further configured to perform initial staining on each staining sub-region by applying different concentrations of a staining agent to each staining sub-region according to a reference light transmittance of lymphocyte tissue in each staining sub-region, and to obtain a color gradient vector of the lymph tissue section; and determine light absorption characteristics of each staining sub-region by using a luminometer, and determine a staining feature of the lymph tissue section based on all the light absorption characteristics and the color gradient vector. In addition, the processing module 202 is further configured to obtain a staining image of the lymph tissue section after staining processing, extract deep features of each staining sub-region from the staining image, and determine a staining loss value of each staining sub-region based on the deep features and the staining feature. The execution module 203 is mainly configured to perform adaptive adjustment on the concentration of the staining agent of each staining sub-region based on the staining loss value.

[0060] In addition, the present application further provides a computer device, which comprises a memory and a processor, the memory stores codes, and the processor is configured to acquire the codes and execute the pathological tissue section staining control method.

[0061] In some embodiments, referring to Figure 5 The figure is a structural schematic diagram of a computer device for implementing the pathological tissue section staining control method according to some embodiments of the present application. The pathological tissue section staining control method in the above embodiments can be implemented by the computer device shown in the figure, which is a computer device for implementing the pathological tissue section staining control method. Figure 5 The computer device 300 comprises at least one processor 301, a communication bus 302, a memory 303, and at least one communication interface 304.

[0062] The processor 301 can be a general central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more circuits for controlling the execution of the pathological tissue section staining control method in the present application.

[0063] The communication bus 302 can be used to transmit information between the above components.

[0064] The memory 303 can be a read-only memory (ROM) or other type of static storage device that can store static information and instructions, a random access memory (RAM), or other type of dynamic storage device that can store information and instructions, and can be an electrically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM), or other optical disk storage, a magneto-optical disk, a magnetic disk or other magnetic storage device, or any other medium capable of storing desired program code in the form of instructions or data structures and that can be accessed by a computer, but is not limited thereto. The memory 303 can exist independently, and is connected to the processor 301 through the communication bus 302. The memory 303 can also be integrated with the processor 301.

[0065] The memory 303 is configured to store program codes for implementing the solutions of the present application, and the processor 301 is configured to execute the program codes stored in the memory 303. The program codes can include one or more software modules. The determination of the pathological tissue slice staining control method in the above-described embodiments can be implemented by the processor 301 and one or more software modules in the program codes in the memory 303.

[0066] The communication interface 304 is configured to communicate with other devices or communication networks, such as an Ethernet, a radio access network (RAN), a wireless local area network (WLAN), and the like, using any transceiver-like mechanism.

[0067] In a specific implementation, as an example, the computer device can include a plurality of processors, each of which can be a single-CPU processor or a multi-CPU processor. The processor herein can refer to one or more devices, circuits, and / or processing cores for processing data (e.g., computer program instructions).

[0068] The computer device described above can be a general-purpose computer device or a special-purpose computer device. In a specific implementation, the computer device can be a desktop computer, a laptop computer, a network server, a personal digital assistant (PDA), a mobile phone, a tablet computer, a wireless terminal device, a communication device, or an embedded device. The embodiments of the present application do not limit the type of the computer device.

[0069] In addition, the present application also provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to implement the pathological tissue section staining control method.

[0070] Although the preferred embodiments of the present application have been described, those skilled in the art who are informed of the basic inventive concept can make additional changes and modifications to the embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of the present application.

[0071] Obviously, those skilled in the art can make various modifications and variations to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalent technologies, the present application also intends to include these modifications and variations.

Claims

1. A method for controlling staining of pathological tissue sections, characterized in that: The steps include: Obtain lymphoid tissue sections to be stained; dividing the lymphoid tissue section into a plurality of staining partitions according to the structural characteristics of lymphocyte tissue in the lymphoid tissue section; applying a dye of different concentrations to each stained section for initial staining based on a baseline transmittance of lymphocyte tissue in each stained section, thereby obtaining a color gradient vector of the lymphoid tissue section; measuring a light absorption characteristic of each stained section using a photometer; and determining a staining characteristic of the lymphoid tissue section based on all light absorption characteristics and the color gradient vector; Acquiring a stained image of the lymphatic tissue section after staining, extracting deep features of each stained partition from the stained image, and determining a staining loss value when staining each stained partition based on the deep features and the staining features; The dye concentration of each dyeing partition is adaptively adjusted based on the dyeing loss value.

2. The method according to claim 1, wherein Dividing the lymphoid tissue section into a plurality of staining partitions according to the structural characteristics of the lymphocyte tissue in the lymphoid tissue section specifically includes: acquiring a full-field image of the lymphoid tissue section; extracting structural features of lymphocyte tissue in the lymphoid tissue section from the full-field image; The lymphatic tissue section is divided into regions based on the structural features to obtain multiple staining partitions.

3. The method according to claim 1, wherein Applying different concentrations of dye to each stained section for initial staining according to the baseline transmittance of the lymphocyte tissue in each stained section, thereby obtaining the color gradient vector of the lymphatic tissue section specifically includes: Constructing a database of mapping relationships between transmittance and dye concentration; Determining the baseline transmittance of the lymphocyte tissue in each stained partition; Performing a matching query in the mapping relationship database according to the reference transmittance to obtain a concentration of the adapted dye for each dyeing partition; The cell tissue in each staining partition is stained based on the concentration of the adapted dye corresponding to each staining partition, thereby obtaining a stained section; A color gradient vector of the lymphatic tissue section is extracted based on the stained section.

4. The method according to claim 1, wherein Determining the staining characteristics of the lymphatic tissue section based on all light absorption characteristics and the color gradient vector specifically includes: Constructing a high-dimensional feature based on all light absorption features and the color gradient vector; The staining features of the lymphatic tissue section are extracted from the high-dimensional features.

5. The method according to claim 1, wherein Extracting the deep features of each stained partition from the stained image specifically includes: Performing partition positioning on all the stained partitions in the stained image to obtain a plurality of sub-image blocks, each sub-image block corresponding to a stained partition; Feature extraction is performed on each sub-image block respectively, thereby obtaining deep features of each stained partition.

6. The method according to claim 1, wherein Adaptively adjusting the dye concentration of each dyeing partition based on the dyeing loss value specifically includes: determining a control error corresponding to the dyeing loss value; Obtaining the concentration of the adapted dye corresponding to each stained partition; The dye concentration of each dyeing partition is updated based on the adapted dye concentration and the control error to achieve adaptive adjustment of the dye concentration.

7. The method according to claim 1, wherein Lymphoid tissue sections were obtained using a standardized pathological tissue section preparation process.

8. A pathological tissue section staining device, comprising a staining control unit, characterized in that: The dyeing control unit comprises: An acquisition module, used for acquiring lymphatic tissue sections to be stained; a processing module, configured to divide the lymphatic tissue section into a plurality of staining partitions according to structural characteristics of lymphocyte tissue in the lymphatic tissue section; The processing module is further configured to apply different concentrations of dye to each stained section for initial staining based on a baseline transmittance of lymphocyte tissue in each stained section, thereby obtaining a color gradient vector of the lymphatic tissue section, measure light absorption characteristics of each stained section using a photometer, and determine a staining characteristic of the lymphatic tissue section based on all light absorption characteristics and the color gradient vector; The processing module is further configured to obtain a stained image after staining the lymphatic tissue section, extract deep features of each stained partition from the stained image, and determine a staining loss value when staining each stained partition based on the deep features and the staining features; An execution module is used to adaptively adjust the dye concentration of each dyeing partition based on the dyeing loss value.

9. A computer device, characterized in that: The computer device includes a memory and a processor, the memory stores codes, and the processor is configured to acquire the codes and execute the pathological tissue section staining control method according to any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the pathological tissue section staining control method according to any one of claims 1 to 7 is implemented.

Images