Dyeing prediction method and system for flow cytometry of cells

Abstract

The embodiment of the invention provides a staining prediction method and system for flow cytometry of cells, and belongs to the technical field of data processing. The method comprises the steps of obtaining historical experiment data, and performing data clustering on the historical experiment data to obtain multiple pieces of sub experiment data; obtaining associated image information corresponding to the sub-experiment data from the historical experiment data; performing edge identification on the associated image information to obtain adherent cell information and clotted cell information corresponding to the sub-experiment data; obtaining to-be-dyed data, and calculating similar information between the to-be-dyed data and the sub-experiment data; performing information fusion on the sub-experimental data according to the similar information to obtain an initial dyeing parameter corresponding to the to-be-dyed data; adjusting the initial dyeing parameters according to the adherent cell information and the clotted cell information to obtain target dyeing parameters corresponding to the to-be-dyed data; and performing dyeing prediction on the to-be-dyed data according to the target dyeing parameter to obtain a dyeing prediction result corresponding to the to-be-dyed data.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a staining prediction method and system for cell flow cytometry. Background Art

[0002] Regulatory B cells (Bregs) are a type of B cells that negatively regulate the body's immune response. They mainly mediate immune control by secreting cytokines, granzymes, and interacting with cell surface molecules. Studies have shown that the increase in the number of Bregs plays an important role in chronic infection and tumor recurrence and metastasis. Therefore, accurately determining the typing of regulatory B cells in pathological samples is of great significance for clinical diagnosis and treatment. In actual operation, when staining regulatory B cells, factors such as dye concentration and staining time may affect the experimental results. If the dye concentration is too high or too low, or the staining time is insufficient, the polyploid peak of DNA cannot be displayed, thereby affecting the accuracy of the data. In addition, other unfavorable factors in the staining process, such as cell adhesion and agglomeration risk, may also interfere with the results. Although staining parameter prediction based on historical data can provide a certain reference, the existing technology still has shortcomings in prediction accuracy and cannot fully consider all potential influences in the staining process of regulatory B cells.

[0003] Therefore, it is urgent to design a technical solution to solve at least one of the above technical problems. Summary of the invention

[0004] The main purpose of the embodiments of the present application is to provide a staining prediction method and system for cell flow cytometry, aiming to solve the problems of insufficient accuracy in staining prediction in related technologies and failure to fully consider the potential impact of regulatory B cell staining process.

[0005] In a first aspect, the present application provides a staining prediction method for cell flow cytometry, comprising:

[0006] Obtaining historical experimental data, and performing data clustering on the historical experimental data to obtain a plurality of sub-experimental data;

[0007] Obtaining associated image information corresponding to the sub-experimental data from the historical experimental data;

[0008] Performing edge recognition on the associated image information to obtain the adherent cell information and the clumping cell information corresponding to the sub-experimental data;

[0009] Obtaining data to be stained, and calculating similarity information between the data to be stained and the sub-experiment data;

[0010] The sub-experiment data are subjected to information fusion according to the similar information to obtain initial staining parameters corresponding to the data to be stained;

[0011] Adjusting the initial staining parameters according to the adherent cell information and the clumping cell information to obtain target staining parameters corresponding to the data to be stained;

[0012] Performing dyeing prediction on the data to be dyed according to the target dyeing parameters to obtain a dyeing prediction result corresponding to the data to be dyed.

[0013] In a second aspect, the present application provides a staining prediction system for cell flow cytometry, comprising:

[0014] A data clustering module is used to obtain historical experimental data and perform data clustering on the historical experimental data to obtain multiple sub-experimental data;

[0015] An image acquisition module, used to obtain associated image information corresponding to the sub-experiment data from the historical experiment data;

[0016] An edge processing module, used for performing edge recognition on the associated image information to obtain the adherent cell information and the clumping cell information corresponding to the sub-experiment data;

[0017] A similarity calculation module, used for obtaining the data to be stained and calculating the similarity information between the data to be stained and the sub-experiment data;

[0018] A parameter determination module, used for fusing the sub-experiment data according to the similar information to obtain initial staining parameters corresponding to the data to be stained;

[0019] A parameter adjustment module, used for adjusting the initial staining parameters according to the adherent cell information and the clumping cell information to obtain target staining parameters corresponding to the data to be stained;

[0020] The dyeing prediction module is used to perform dyeing prediction on the data to be dyed according to the target dyeing parameters to obtain a dyeing prediction result corresponding to the data to be dyed.

[0021] In a third aspect, an embodiment of the present application further provides a terminal device, comprising a processor, a memory, a computer program stored in the memory and executable by the processor, and a data bus for realizing connection and communication between the processor and the memory, wherein when the computer program is executed by the processor, the steps of any one of the staining prediction methods for cell flow cytometry detection provided in the specification of the present invention are implemented.

[0022] In a fourth aspect, an embodiment of the present application also provides a storage medium for computer-readable storage, characterized in that the storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement the steps of any cell flow cytometry staining prediction method provided in the specification of the present invention.

[0023] The present application embodiment provides a staining prediction method and system for cell flow detection, the method comprising: collecting and collating historical experimental data and performing data clustering analysis on the historical experimental data, which can be divided into multiple sub-experimental data. Each sub-experimental data set represents a class of data with similar characteristics. Next, the associated image information associated with each sub-experimental data is extracted from the historical experimental data. By edge recognition of the associated image information, the adherent cell information and the agglomerated cell information related to each sub-experimental data can be obtained in detail. This information is crucial for subsequent analysis and prediction. Then, the similar information between the data to be stained and the sub-experimental data is calculated. This step can help identify the correlation between the data to be stained and the historical experimental data. Based on these similar information, the sub-experimental data are subjected to information fusion. During the information fusion process, the weights of different sub-experimental data are comprehensively considered, thereby improving the scientificity and rationality of the initial setting of the staining parameters. In this way, more representative and accurate initial staining parameters can be obtained. After obtaining the initial staining parameters, the initial staining parameters are adjusted according to the previously extracted adherent cell information and agglomerated cell information. This step aims to optimize the initial staining parameters according to the cell type and distribution characteristics to obtain the target staining parameters. This adjustment helps to achieve a more precise staining effect and ensure the accuracy of the staining process. Finally, based on the adjusted target staining parameters, the staining data is predicted. This step uses the optimized parameters to generate staining prediction results, so that the final staining of the staining data can be accurately predicted. In this way, experimenters can obtain more reliable and accurate target staining results, effectively improving the overall efficiency and credibility of the results. It also solves the problem of insufficient accuracy in staining prediction in related technologies and the inability to fully consider the potential impact of regulatory B cell staining. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0025] Figure 1 A schematic diagram of a staining prediction method for cell flow cytometry provided in an embodiment of the present application;

[0026] Figure 2 A schematic diagram of the module structure of a staining prediction system for cell flow cytometry detection provided in an embodiment of the present application;

[0027] Figure 3 A schematic block diagram of the structure of a terminal device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0028] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0029] The flowcharts shown in the accompanying drawings are only examples and do not necessarily include all the contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps may also be decomposed, combined or partially merged, so the actual execution order may change according to actual conditions.

[0030] It should be understood that the terms used in this specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. As used in the specification of the present invention and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include plural forms.

[0031] The present application embodiment provides a method and system for predicting staining of cell flow cytometry. The method and system for predicting staining of cell flow cytometry can be applied to a terminal device, which can be an electronic device such as a tablet computer, a laptop computer, a desktop computer, a personal digital assistant, and a wearable device. The terminal device can be a server or a server cluster.

[0032] Some embodiments of the present invention are described in detail below in conjunction with the accompanying drawings. In the absence of conflict, the following embodiments and features in the embodiments can be combined with each other.

[0033] Please refer to Figure 1 , Figure 1 A schematic flow chart of a staining prediction method for cell flow cytometry provided in an embodiment of the present application.

[0034] like Figure 1 As shown, the staining prediction method for cell flow cytometry detection includes steps S101 to S107.

[0035] Step S101: Obtain historical experimental data, and perform data clustering on the historical experimental data to obtain a plurality of sub-experimental data.

[0036] Exemplarily, B cells are isolated from peripheral blood, spleen or lymph nodes. Methods such as density gradient centrifugation or magnetic bead sorting can be used to obtain a cell colony rich in B cells. Cell counting plates or automatic cell counters are used to measure the concentration of cells for subsequent staining. Suitable fluorescent-labeled antibodies are selected for regulatory B cell-specific markers (e.g., CD19, CD24, CD38, CD5, CD1d, etc.). These antibodies can help identify the surface features of regulatory B cells.

[0037] Exemplarily, single staining or multiple staining methods can be used for cell staining. Multiple staining can detect multiple cell surface markers simultaneously by fluorescently labeled antibodies. The selected fluorescently labeled antibody is diluted according to the manufacturer's instructions according to an appropriate dilution ratio. Cell pretreatment can be performed with an appropriate cell culture medium or buffer to reduce nonspecific binding. The cell suspension is mixed with the diluted antibody and gently shaken. The mixture is incubated at a preset temperature and in the dark for a preset time to complete the staining of regulatory B cells.

[0038] Exemplarily, a staining experiment of regulatory B cells is set according to parameters such as cell concentration, staining time, and staining temperature to obtain initial experimental data, and the initial experimental data is screened according to the staining results to obtain historical experimental data, that is, the historical experimental data is experimental data that can obtain staining results that meet preset conditions under the staining parameters.

[0039] Exemplarily, a corresponding clustering algorithm is selected according to the dyeing parameters. The clustering algorithm includes but is not limited to K-means, hierarchical clustering or k-nearest neighbor clustering, etc. The historical experimental data is clustered according to the selected clustering algorithm to obtain a plurality of sub-experimental data.

[0040] In some embodiments, the historical experimental data includes at least a first experimental object and a second experimental object, and the data clustering of the historical experimental data to obtain multiple sub-experimental data includes: obtaining a first staining result of the first experimental object in a first time series from the historical experimental data and obtaining a second staining result of the second experimental object in a second time series from the historical experimental data; performing data fitting on the first experimental object according to the first time series and the first staining result to obtain a first fitting result corresponding to the first experimental object; performing data fitting on the second experimental object according to the second time series and the second staining result to obtain a second fitting result corresponding to the second experimental object; performing data clustering on the first experimental object and the second experimental object according to the first fitting result and the second fitting result to obtain multiple sub-experimental data.

[0041] Exemplarily, the historical experimental data includes multiple groups of experimental data, including at least a first experimental object and a second experimental object, and when collecting the historical experimental data, the staining results of different staining times under the same parameters are obtained in real time. Subsequently, the first staining result of the first experimental object in the first time series is obtained from the historical experimental data according to the time series data, and the second staining result of the second experimental object in the second time series is obtained according to the time series data.

[0042] Exemplarily, a staining quality assessment is performed on the first staining result to obtain a first staining quality corresponding to the first staining result, and then the first staining parameter and the first staining quality are fitted according to a first time series to obtain a first fitting result of the first experimental object.

[0043] Exemplarily, a staining quality assessment is performed on the second staining result to obtain a second staining quality corresponding to the second staining result, and then the second staining parameter and the second staining quality are fitted according to a second time series to obtain a second fitting result of the second experimental object.

[0044] Exemplarily, a first staining trend corresponding to the first fitting result and a second staining trend corresponding to the second fitting result are determined, and then data clustering is performed on the first experimental object and the second experimental object according to the first staining trend and the second staining trend, thereby obtaining multiple sub-experimental data.

[0045] For example, the trend of change of dyeing is extracted from the first fitting result. This may include the slope, shape, change pattern, etc. of the first fitting result. Determine the characteristics of the dyeing trend, such as linear increase, nonlinear change, periodic fluctuation, etc. These trends can be identified by analyzing the parameters of the fitting results or visualizing the fitting curve. The trend of change of dyeing is extracted from the second fitting result, which is similar to the step of analyzing the first fitting result. Thus, a feature vector is constructed according to the first dyeing trend and the second dyeing trend. Each experimental object can be represented by a vector containing its dyeing trend characteristics. For example, the slope, intercept, change pattern, etc. of the trend can be converted into numerical features. And ensure that all feature vectors are on the same scale for effective clustering. Thus, the experimental objects are divided into multiple clusters based on the similarity of the dyeing trends of different experimental objects under the same time series, and then multiple sub-experimental data are obtained.

[0046] Specifically, by determining the dyeing trends corresponding to the first fitting result and the second fitting result, the pattern of dyeing changes over time or conditions can be clearly understood. This helps to understand the changes in dyeing behavior under different experimental conditions. Experimental objects can be clustered according to dyeing trends, and experimental objects with similar trends can be grouped in the same group. Cluster analysis can then reveal the potential relationships and rules between different experimental objects. This provides support for subsequent dyeing parameter settings.

[0047] Step S102: Obtain associated image information corresponding to the sub-experiment data from the historical experiment data.

[0048] Exemplarily, when collecting historical experimental data, the experimental object is imaged by an imaging device to obtain experimental image information, and then after data clustering analysis, the associated image information corresponding to each sub-experimental data is obtained from the experimental image information of the historical experimental data.

[0049] Step S103: performing edge recognition on the associated image information to obtain the adherent cell information and clumping cell information corresponding to the sub-experimental data.

[0050] Exemplarily, the associated image information includes the wall-attached image of the culture dish and the corresponding top-view image of the culture dish collected during the sub-experiment data experiment.

[0051] Exemplarily, denoising techniques (such as Gaussian blur, median filtering, etc.) are used to reduce the noise in the image for the wall-attached image and the top-view image, respectively, so that the accuracy of edge recognition can be improved, and then the contrast of the image can be adjusted to make the cell edge more obvious. Histogram equalization or other contrast adjustment methods can be used. Thus, edge detection algorithms such as Sobel operator, Canny edge detection, Laplacian operator and other algorithms are used to identify the edges of cells in the wall-attached image and the top-view image, thereby obtaining the first edge information corresponding to the wall-attached image and the second edge information corresponding to the top-view image.

[0052] For example, the cell regions in contact with the bottom are identified by analyzing the first edge information. Adherent cells usually have obvious borders and bottom contact areas, which can be identified by analyzing edge and shape features, and then the features of the adherent cells, such as the shape, size, position, etc. of the cells, are extracted, thereby obtaining the adherent cell information corresponding to the adherent image.

[0053] Exemplarily, the cell regions that are in contact or overlap with each other in the second edge information are analyzed to identify cell groups or clustering regions. Clustered cells are usually in the form of multiple cells closely arranged together. The characteristics of the clustered cells are extracted, such as the density, shape, and region of the clusters. Thus, the clustered cell information corresponding to the top view image is obtained.

[0054] In some embodiments, the associated image information includes an initial adherent image and a plane image, and the edge recognition of the associated image information to obtain the adherent cell information and clumping cell information corresponding to the sub-experimental data includes: registering and fusing any two sub-images in the initial adherent image to obtain a connection sequence corresponding to the initial adherent image; splicing the sub-images according to the connection sequence to obtain a target adherent image; performing edge recognition on the target adherent image to obtain the adherent cell information corresponding to the target adherent image; performing edge recognition on the plane image to obtain an edge image corresponding to the plane image, and performing image segmentation on the edge image to obtain a segmented object; and performing density calculation based on the segmented object to obtain the clumping cell information corresponding to the plane image.

[0055] Exemplarily, the associated image information includes an initial wall-attached image and a plane image. The initial wall-attached image is an image of the culture dish collected during the experiment of the sub-experiment data, and the plane image is a top-view image corresponding to the culture dish.

[0056] Exemplarily, the initial wall-attached image includes multiple sub-images corresponding to the culture dish at different angles, and then the wall-attached image of the culture dish around the circle can be obtained by stitching the sub-images, and then any two sub-images in the initial wall-attached image are registered and fused using feature point matching, image transformation model, etc. to obtain the connection order corresponding to the initial wall-attached image, and then the sub-images are stitched according to the connection order to obtain the target wall-attached image.

[0057] Exemplarily, the registered sub-images are fused according to the connection order, and then during the stitching process, attention is paid to processing the smooth transition of the image boundary to ensure the coherence of the final stitched image, thereby obtaining a complete target wall-attached image.

[0058] Exemplarily, the cell edges in the target adherent image are identified according to an edge detection algorithm (such as Canny, Sobel, etc.), and then the edge detection algorithm is applied to the target adherent image to generate an adherent edge image, thereby analyzing the adherent edge image and extracting the adherent cell information of the adherent cells. The adherent cell information may include the shape, boundary, distribution, etc. of the cells.

[0059] Exemplarily, similar to the edge recognition of the target adherent image, a suitable edge detection algorithm (such as Canny, Sobel, etc.) is selected to process the plane image. The edge detection algorithm is applied to the plane image to generate an edge image. An appropriate image segmentation method (such as threshold segmentation, region growing, etc.) is selected to segment the objects in the edge image. The edge image is segmented into multiple objects using the selected segmentation method, thereby calculating the number and area of ​​the segmented objects to evaluate the density of the clumping cells. The density of the clumping cells in the plane image is then calculated based on the statistical data of the segmented objects (such as the total number of objects, area, etc.), and the density of the clumping cells is then determined as the clumping cell information.

[0060] Specifically, each sub-image in the initial adherent image is first accurately registered. This step ensures the spatial correspondence between the sub-images and lays the foundation for subsequent stitching. The registered sub-images are seamlessly stitched to generate a complete and continuous target adherent image. This integration method not only provides a global view of the cell distribution in the experimental area, but also helps researchers grasp the overall layout of the cells from a macroscopic perspective. In the stitching process, attention is paid to reducing image distortion and stitching errors. Through sophisticated algorithm processing, the stitched target adherent image is of high quality and clear details, laying a solid foundation for the subsequent in-depth analysis. The target adherent image is then processed by edge recognition. This step can accurately depict the outline and key features of the adherent cells, allowing researchers to observe and analyze the morphology, spatial distribution and their relationship in more detail. In addition, the adherent cell information extracted by edge recognition can accurately count the number of cells and deeply analyze their distribution patterns. For planar images, we also use edge recognition technology to clearly define cell clusters and other image areas. This fine segmentation process enables more accurate identification and labeling of clumped cells, providing a reliable data basis for subsequent analysis, thereby further calculating the density of these segmented objects. This data not only reveals the number of clumped cells, but also reflects their distribution density. Through this quantitative information, researchers can gain a deeper understanding of the characteristics and behavior patterns of cell clumping, providing strong data support for subsequent scientific research.

[0061] In some embodiments, the stitching of the sub-images according to the connection order to obtain a target wall-attached image includes: obtaining adjacent left and right images from the sub-images according to the connection order; cutting the left image to obtain a first registration area and cutting the right image to obtain a second registration area according to the registration fusion result; calculating a first gradient corresponding to the first registration area and calculating a second gradient corresponding to the second registration area; determining a corresponding gradient difference value according to the first gradient and the second gradient; performing brightness difference analysis on the first registration area and the second registration area to obtain a brightness difference value; fusing the gradient difference value and the brightness difference value to determine a corresponding registration difference value between the first registration area and the second registration area; adjusting the first registration area and the second registration area according to the registration difference value to obtain a stitching rule corresponding to the left image and the right image; determining the target wall-attached image according to the stitching rule and the connection order; wherein the registration difference value is obtained according to the following formula:

[0062]

[0063] Wherein, c represents the registration difference, w1 represents the weight information corresponding to the gradient difference value, w2 represents the weight information corresponding to the brightness difference value, L x1 represents the gradient of the left image in the x direction, L y1 represents the gradient of the left image in the y direction, L x2 represents the gradient of the right image in the x direction, L y2 Represents the gradient of the right image in the y direction, img 1k represents the brightness information of the left image in k dimension, img 2k represents the brightness information of the right image in the k-dimensional space, n represents the number of dimensional divisions corresponding to the left image or the number of dimensional divisions corresponding to the right image, q k Represents the weight information under the kth dimension.

[0064] Exemplarily, according to the previously defined connection order, the adjacent left image and right image in the initial wall-attached image are determined. According to the result of the registration fusion, the first registration region in the left image is determined, and the first registration region is the part to be compared with the second registration region of the right image. Similarly, the second registration region in the right image is obtained, and the second registration region is the part to be compared with the first registration region of the left image.

[0065] Exemplarily, the first registration region is extracted from the left image, and further gradient calculation and brightness analysis are performed. Similarly, the second registration region is extracted from the right image, and further gradient calculation and brightness analysis are performed.

[0066] For example, a gradient calculation method (such as Sobel operator) is used to calculate a first gradient of the first registration area. The first gradient represents the intensity and direction of the brightness change in the first registration area. Similarly, the second gradient of the second registration area is calculated using the same method. The second gradient represents the intensity and direction of the brightness change in the second registration area.

[0067] Exemplarily, a difference calculation is performed between the first gradient of the first registration area and the second gradient of the second registration area to obtain a gradient difference value.

[0068] Exemplarily, a suitable brightness difference analysis method is selected to calculate the brightness difference value between the first registration area and the second registration area under different chromaticities, which helps to evaluate the degree of brightness matching between the two areas. Thus, the gradient difference value and the brightness difference value are combined to determine the registration difference between the two registration areas. This difference is used to evaluate the registration accuracy between the areas. The registration difference is obtained according to the following formula:

[0069]

[0070] Among them, c represents the registration difference, w1 represents the weight information corresponding to the gradient difference value, w2 represents the weight information corresponding to the brightness difference value, and L x1 Represents the gradient of the left image in the x direction, L y1 Represents the gradient of the left image in the y direction, L x2 represents the gradient of the right image in the x direction, L y2 Represents the gradient of the right image in the y direction, img 1k Represents the brightness information of the left image in dimension k, img 2k represents the brightness information of the right image in k dimensions, n represents the number of dimensional divisions corresponding to the left image or the number of dimensional divisions corresponding to the right image, q k Represents the weight information under the kth dimension.

[0071] For example, different chromaticities or different dimensions include RGB dimension or HIS dimension, etc., then the brightness difference between the first registration area and the second registration area in different dimensions is calculated respectively, and then the weight information in different dimensions is set according to historical experience, and then the brightness difference between the first registration area and the second registration area in multiple dimensions is fused according to the weight information to obtain the brightness difference value. Then the gradient of the first registration area in the x direction and the gradient in the y direction in the left image are calculated respectively, and then the gradient of the first registration area in the x direction and the gradient in the y direction are fused to obtain the first gradient. Similarly, the gradient of the second registration area in the x direction and the gradient in the y direction in the right image are calculated respectively, and then the gradient of the second registration area in the x direction and the gradient in the y direction are fused to obtain the second gradient, and the first gradient and the second gradient of the first registration area and the second registration area are calculated by difference to obtain the gradient difference value. Finally, the weight information corresponding to the gradient difference value and the brightness difference value is determined according to expert experience, and then the registration difference between the first registration area and the second registration area is determined according to different weight information by fusion of the gradient difference value and the brightness difference value.

[0072] Exemplarily, the calculated registration difference is compared with a preset value. When the registration difference is less than or equal to the preset value, the first registration area and the second registration area are directly aligned. Then, after weighted summation of the first registration area and the second registration area, the weighted result is determined as the target registration area after the first registration area and the second registration area correspond. Then, the stitching rule corresponding to the left image and the right image is to cut the first registration area from the left image to obtain the first image, and to cut the second registration area from the right image to obtain the second image. Then, the first image, the target registration area, and the second image are stitched in order. When the registration difference is greater than the preset value, the position of the first registration area in the left image is adjusted to regain the first registration area, and the position of the second registration area in the right image is adjusted to regain the second registration area, and then the registration difference between the first registration area and the second registration area is recalculated until the registration difference is less than or equal to the preset value.

[0073] Exemplarily, all the splicing rules between all the left images and the right images are obtained in sequence according to the connection order, and then all the sub-images are spliced ​​according to all the splicing rules to obtain the target wall-attached image.

[0074] Specifically, by performing detailed cutting, gradient and brightness analysis on adjacent left and right images, the registration difference between the two areas can be accurately determined. This precise analysis ensures accurate alignment of the images and reduces errors in the stitching process. By calculating the gradient difference and brightness difference and using these difference values ​​to adjust the image, it helps to reduce the stitching error of the stitching boundary and improve the coherence of the stitching area. Thus, the gradient difference value and the brightness difference value are fused to determine the registration difference, providing a comprehensive standard for evaluating the quality of image registration, making image stitching more accurate and reliable. The resulting target adherent image provides a comprehensive view of the entire experimental area, which helps to evaluate and analyze the cell distribution as a whole. It provides solid data support for further scientific research.

[0075] In some embodiments, the obtaining of the clumping cell information corresponding to the plane image by density calculation based on the segmented object includes: determining a segmentation window, sliding the segmentation window in the plane image to obtain a window image; obtaining an associated object corresponding to the window image from the segmented image according to the plane image; obtaining density information corresponding to the window image by density calculation based on the associated object; when the density information is greater than or equal to a preset value, obtaining an area value by area calculation for the associated object in the window image; and obtaining the clumping cell information by screening out the associated object according to the area value.

[0076] For example, by defining a segmentation window in a planar image and sliding it, the entire image can be systematically covered. This approach ensures that each region is analyzed separately, improving the comprehensiveness and detail of the segmentation. Processing the window image corresponding to each sliding window can locally analyze features in the image, helping to capture details and changes.

[0077] Exemplarily, the associated object corresponding to the window image is extracted from the segmented image, that is, the edge information corresponding to the window image is obtained from the segmented image.

[0078] Exemplarily, by performing density calculation on the associated objects in the window image, the density information of the cells in each segmented window is obtained, and the density information provides quantitative data on the cell distribution density, which is crucial for determining whether there are cell clusters in the region. When the density information is greater than or equal to the preset value, the area of ​​the associated objects in the window image is further calculated. This helps to screen out areas with higher density, thereby locating possible cell clusters. Thus, calculating the area value can help determine the actual size of each cluster, which is very important for further analyzing the characteristics and behavior of cell clusters.

[0079] Illustratively, by screening out related objects based on their area values, objects that do not meet the standards can be excluded to obtain more accurate clumping cell information, thereby ensuring that the extracted clumping cell information is more accurate and meaningful, and providing a reliable data basis for subsequent analysis.

[0080] Specifically, through refined segmentation, precise extraction of associated objects, density calculation, and area analysis, cell clustering can be more accurately identified and quantified. This not only optimizes the data processing process, but also enhances the understanding of cell distribution and clustering, providing strong support for scientific research and experiments.

[0081] Step S104, obtaining the data to be stained, and calculating the similarity information between the data to be stained and the sub-experiment data.

[0082] Exemplarily, B cells are isolated from peripheral blood, spleen or lymph nodes. Density gradient centrifugation or magnetic bead sorting or the like can be used to obtain a cell population rich in B cells, and then the separation result is determined as the data to be stained.

[0083] Exemplarily, parameters such as cell concentration and cell type corresponding to the data to be stained are obtained, and then similarity information between the data to be stained and the sub-experimental data is calculated.

[0084] For example, the data to be stained is represented by vectors to obtain a first vector, and each data in the sub-experimental data is represented by vectors and then fused to obtain a second vector, and then similarity information is calculated based on the first vector and the second vector using cosine similarity.

[0085] Step S105 , fusing the sub-experiment data according to the similar information to obtain initial staining parameters corresponding to the data to be stained.

[0086] Exemplarily, the sub-experimental data closest to the data to be stained are screened out from multiple sub-experimental data based on similar information, and then the regression function between the staining data and the staining parameters is determined based on the sub-experimental data, and then the data to be stained is input into the regression function to obtain the initial staining parameters corresponding to the data to be stained.

[0087] In some embodiments, the sub-experimental data are subjected to information fusion according to the similarity information to obtain initial staining parameters corresponding to the data to be stained, including: determining the similarity between the data to be stained and the sub-experimental data according to the similarity information; calculating the probability value corresponding to the data to be stained belonging to the sub-experimental data according to the similarity; obtaining predicted staining parameters corresponding to the staining prediction of the data to be stained under the sub-experimental data; and performing weighted summation of the predicted staining parameters according to the probability value to obtain the initial staining parameters corresponding to the data to be stained.

[0088] Exemplarily, similar information corresponding to parameters such as cell concentration and cell type of the data to be stained is weightedly fused to obtain the similarity between the data to be stained and the sub-experimental data, and then based on the calculated similarity, a Gaussian mixture model, Bayesian method, etc. are used to calculate the probability value that the data to be stained belongs to each sub-experimental data.

[0089] Exemplarily, a corresponding parameter prediction model is trained for each sub-experimental data, and the parameter prediction model is a data regression model. The preprocessed data to be stained is taken as input and input into the parameter prediction model corresponding to each sub-experimental data respectively, and then the data to be stained is passed to the parameter prediction model and predicted to obtain the predicted staining parameters corresponding to each parameter prediction model.

[0090] Exemplarily, the predicted staining parameters of each sub-experimental data are weighted using the calculated probability value. Specifically, the predicted staining parameters of each sub-experimental data are multiplied by the corresponding probability value, and all weighted results are added together to obtain the final initial staining parameters, that is, the initial staining parameters are calculated based on the predicted staining parameters of each sub-experimental data and their probability values.

[0091] Specifically, by integrating the predictions of multiple sub-experimental data, the possible deviations of a single model are reduced, thereby improving the robustness and adaptability of the overall model. This process provides flexible adaptation space for changes in different sub-experimental data. With the addition of new sub-experimental data, the prediction parameters can be dynamically adjusted to keep the model updated and timely, which helps to improve the quality and reliability of staining predictions, so that more satisfactory results can be obtained in practical applications.

[0092] Step S106: adjusting the initial staining parameters according to the adherent cell information and the clumping cell information to obtain target staining parameters corresponding to the data to be stained.

[0093] Exemplarily, the sub-experiment data closest to the characteristics of the data to be stained is found from the multiple sub-experiment data according to the similar information, so as to obtain the adherent cell information and clumping cell information corresponding to the closest sub-experiment data for subsequent analysis.

[0094] Exemplarily, the selected sub-experimental data are analyzed in detail to extract the information of the adherent cells and the clumping cells. This information may include the time point of cell attachment, the type of cells attached, the characteristics of clumping, the time point of clumping, etc. Based on the information in the selected sub-experimental data, a prediction model is established or a statistical method is used to predict the time of the possible adhesion effect and clumping effect of the data to be stained. Factors such as the growth rate of the cells, the tendency to adhere to the wall, and the tendency of clumping are considered.

[0095] Exemplarily, a dynamic culture system, such as a shaking table or a stirring culture device, is set up according to the prediction results. The system can operate at the predicted time points of adhesion and agglomeration effects to help cells maintain a suspended state. Thus, the shaking table or stirring device is started at the predicted time point, and the suspension state of the cells is continuously monitored to ensure that the cells do not adhere to the wall or agglomerate during the culture process. The trigger time of the dynamic culture system is integrated into the initial staining parameters. Consider the adjustment time points that may be needed during the staining process to optimize the staining effect. Thus, the target staining parameters are updated after considering the intervention of the dynamic culture system. These parameters should include the time points of dynamic culture and their effects on the staining effect.

[0096] For example, by selecting the most similar sub-experiment data from multiple sub-experiment data, combining the information of adherent cells and clumping cells for prediction, and optimizing the cell culture using a dynamic culture system, the adhesion and clumping phenomena can be effectively reduced and the staining effect can be improved. Finally, incorporating the intervention time of the dynamic culture system into the initial staining parameters helps achieve the optimal target staining parameters and improves the accuracy and effect of the overall experiment.

[0097] In some embodiments, the adjusting the initial staining parameters according to the adherent cell information and the clumping cell information to obtain the target staining parameters corresponding to the data to be stained includes: performing weighted summation of the adherent cell information according to the probability value to obtain the target adherence information corresponding to the data to be stained; performing weighted summation of the clumping cell information according to the probability value to obtain the target clumping information corresponding to the data to be stained; and adjusting the initial staining parameters according to the target adherence information and the target clumping information to obtain the target staining parameters corresponding to the data to be stained.

[0098] Exemplarily, the adherent cell information corresponding to each sub-experimental data is weighted according to the previously calculated probability value. This means that the adherent cell information provided by more similar sub-experimental data will be given a higher weight. All weighted adherent cell information is summed to obtain the target adherent cell information corresponding to the data to be stained. This target adherent cell information is a prediction result that integrates the influence of all similar sub-experimental data.

[0099] For example, similar to the adherent cell information, all relevant clumping cell information is collated, including characteristics such as the size, number, and compactness of the clumping. Each clumping cell information is weighted using the same probability value to ensure that similar sub-experimental data have a greater influence on the results. All weighted clumping cell information is summed to obtain the target clumping information corresponding to the data to be stained. This information reflects the expected situation of cell clumping under similar conditions.

[0100] Exemplarily, the target adhesion information and target agglomeration information obtained by weighted summation are carefully analyzed to understand their possible impact on the staining process. Based on the results of the analysis, determine how to adjust the initial staining parameters to cope with the expected adhesion and agglomeration conditions. For example, if more adherent cells are expected, it may be necessary to increase the concentration of the stain or extend the staining time. Apply the determined adjustment strategy to the initial staining parameters and make corresponding adjustments. This may include changing parameters such as the type, concentration, staining time, and temperature of the stain. The adjusted staining parameters are the target staining parameters corresponding to the data to be stained. These parameters should be able to better adapt to the expected cell adhesion and agglomeration conditions, thereby improving the quality and effect of staining.

[0101] In addition, information such as the time point of cell adhesion corresponding to the target adhesion information and the time point of cell agglomeration corresponding to the target agglomeration information can also be extracted, and then a dynamic culture system such as a shaker or stirring culture device can be set according to the prediction results. The system can operate at the predicted time point of cell adhesion and the time point of cell agglomeration to help cells maintain a suspended state. Thus, the shaker or stirring device is started at the predicted time point, and the suspension state of the cells is continuously monitored to ensure that the cells do not adhere to the wall or agglomerate during the culture process. The trigger time of the dynamic culture system is integrated into the initial staining parameters. Consider the adjustment time points that may be needed during the staining process to optimize the staining effect. Thus, the target staining parameters are updated after considering the intervention of the dynamic culture system.

[0102] Specifically, by comprehensively considering the information of adherent cells and clumping cells, the staining parameters can be adjusted more accurately, making the staining process more consistent with the actual cell state. This adjustment can significantly improve the accuracy and consistency of the staining results. Combining the weighted summation information can reduce the staining errors caused by changes in experimental conditions or differences in cell behavior, thereby improving the staining effect of the data to be stained.

[0103] Step S107 , performing dyeing prediction on the data to be dyed according to the target dyeing parameters to obtain a dyeing prediction result corresponding to the data to be dyed.

[0104] Exemplarily, the type of staining prediction model is determined, which may include a mathematical model, a statistical model, or a machine learning model. The model should be able to predict the staining effect based on the target staining parameters. The input parameters of the model are set according to the experimental requirements and the target staining parameters. These parameters may include staining agent concentration, staining time, temperature, time point of cell agglomeration, time point of cell adhesion, etc. Thereby using the staining prediction model to calculate the staining prediction result.

[0105] Exemplarily, the process of determining the type of dyeing prediction model, setting the input parameters of the model, and using the model to calculate the dyeing prediction results includes selecting a suitable model type, defining and accurately inputting parameters, applying the model to make predictions, and analyzing and verifying the prediction results. This ensures that the prediction results accurately reflect the actual experimental conditions, thereby optimizing the dyeing process.

[0106] The present application embodiment provides a staining prediction method and system for cell flow detection, the method comprising: collecting and collating historical experimental data and performing data clustering analysis on the historical experimental data, which can be divided into multiple sub-experimental data. Each sub-experimental data set represents a class of data with similar characteristics. Next, the associated image information associated with each sub-experimental data is extracted from the historical experimental data. By edge recognition of the associated image information, the adherent cell information and the agglomerated cell information related to each sub-experimental data can be obtained in detail. This information is crucial for subsequent analysis and prediction. Then, the similar information between the data to be stained and the sub-experimental data is calculated. This step can help identify the correlation between the data to be stained and the historical experimental data. Based on these similar information, the sub-experimental data are subjected to information fusion. During the information fusion process, the weights of different sub-experimental data are comprehensively considered, thereby improving the scientificity and rationality of the initial setting of the staining parameters. In this way, more representative and accurate initial staining parameters can be obtained. After obtaining the initial staining parameters, the initial staining parameters are adjusted according to the previously extracted adherent cell information and agglomerated cell information. This step aims to optimize the initial staining parameters according to the cell type and distribution characteristics to obtain the target staining parameters. This adjustment helps to achieve a more precise staining effect and ensure the accuracy of the staining process. Finally, based on the adjusted target staining parameters, the staining data is predicted. This step uses the optimized parameters to generate staining prediction results, so that the final staining of the staining data can be accurately predicted. In this way, experimenters can obtain more reliable and accurate target staining results, effectively improving the overall efficiency and credibility of the results. It also solves the problem of insufficient accuracy in staining prediction in related technologies and the inability to fully consider the potential impact of regulatory B cell staining.

[0107] See also Figure 2 , Figure 2A staining prediction system 200 for cell flow cytometry detection provided in an embodiment of the present application includes a data clustering module 201, an image acquisition module 202, an edge processing module 203, a similarity calculation module 204, a parameter determination module 205, a parameter adjustment module 206, and a staining prediction module 207, wherein the data clustering module 201 is used to obtain historical experimental data and perform data clustering on the historical experimental data to obtain multiple sub-experimental data; the image acquisition module 202 is used to obtain associated image information corresponding to the sub-experimental data from the historical experimental data; the edge processing module 203 is used to perform edge recognition on the associated image information to obtain the The attached cell information and the clumping cell information corresponding to the sub-experimental data are as follows: a similarity calculation module 204 is used to obtain the data to be stained and calculate the similarity information between the data to be stained and the sub-experimental data; a parameter determination module 205 is used to fuse the sub-experimental data according to the similarity information to obtain the initial staining parameters corresponding to the data to be stained; a parameter adjustment module 206 is used to adjust the initial staining parameters according to the attached cell information and the clumping cell information to obtain the target staining parameters corresponding to the data to be stained; a staining prediction module 207 is used to perform staining prediction on the data to be stained according to the target staining parameters to obtain the staining prediction results corresponding to the data to be stained.

[0108] In some implementations, the historical experimental data includes at least a first experimental object and a second experimental object. In the process of clustering the historical experimental data to obtain a plurality of sub-experimental data, the data clustering module 201 executes:

[0109] Obtaining a first staining result of the first experimental object in a first time series from the historical experimental data and obtaining a second staining result of the second experimental object in a second time series from the historical experimental data;

[0110] Performing data fitting on the first experimental object according to the first time series and the first staining result to obtain a first fitting result corresponding to the first experimental object;

[0111] Performing data fitting on the second experimental object according to the second time series and the second staining result to obtain a second fitting result corresponding to the second experimental object;

[0112] Data clustering is performed on the first experimental object and the second experimental object according to the first fitting result and the second fitting result to obtain a plurality of sub-experimental data.

[0113] In some embodiments, the associated image information includes an initial adherent image and a plane image, and the edge processing module 203 performs, during the process of performing edge recognition on the associated image information to obtain the adherent cell information and clumping cell information corresponding to the sub-experiment data:

[0114] Performing registration and fusion on any two sub-images in the initial wall-attached image to obtain a connection order corresponding to the initial wall-attached image;

[0115] splicing the sub-images according to the connection order to obtain a target wall-attached image;

[0116] Performing edge recognition on the target adherent image to obtain the adherent cell information corresponding to the target adherent image;

[0117] Performing edge recognition on the plane image to obtain an edge image corresponding to the plane image, and performing image segmentation on the edge image to obtain a segmented object;

[0118] Density calculation is performed according to the segmented object to obtain the agglomerated cell information corresponding to the plane image.

[0119] In some implementations, the edge processing module 203 performs, during the process of stitching the sub-images according to the connection order to obtain the target wall-attached image:

[0120] Obtaining adjacent left and right images from the sub-images according to the connection order;

[0121] According to the registration fusion result, the left image is cut to obtain a first registration area, and the right image is cut to obtain a second registration area;

[0122] Calculating a first gradient corresponding to the first registration area and calculating a second gradient corresponding to the second registration area;

[0123] Determine a corresponding gradient difference value according to the first gradient and the second gradient;

[0124] Performing brightness difference analysis on the first registration area and the second registration area to obtain a brightness difference value;

[0125] fusing the gradient difference value and the brightness difference value to determine a corresponding registration difference value between the first registration area and the second registration area;

[0126] Adjusting the first registration area and the second registration area according to the registration difference to obtain a splicing rule corresponding to the left image and the right image;

[0127] Determine the target wall-attached image according to the splicing rule and the connection order;

[0128] The registration difference is obtained according to the following formula:

[0129]

[0130] Wherein, c represents the registration difference, w1 represents the weight information corresponding to the gradient difference value, w2 represents the weight information corresponding to the brightness difference value, L x1 represents the gradient of the left image in the x direction, L y1 represents the gradient of the left image in the y direction, L x2 represents the gradient of the right image in the x direction, L y2 Represents the gradient of the right image in the y direction, img 1k represents the brightness information of the left image in k dimension, img 2k represents the brightness information of the right image in the k-dimensional space, n represents the number of dimensional divisions corresponding to the left image or the number of dimensional divisions corresponding to the right image, q k Represents the weight information under the kth dimension.

[0131] In some embodiments, the edge processing module 203 performs, during the process of performing density calculation according to the segmented object to obtain the agglomerated cell information corresponding to the plane image:

[0132] Determine a split window, slide the split window in the plane image, and obtain a window image;

[0133] Obtaining an associated object corresponding to the window image from the segmented image according to the plane image;

[0134] Performing density calculation according to the associated object to obtain density information corresponding to the window image;

[0135] When the density information is greater than or equal to a preset value, performing area calculation on the associated object in the window image to obtain an area value;

[0136] The associated objects are screened out according to the area values ​​to obtain the clumping cell information.

[0137] In some implementations, the parameter determination module 205 performs, during the process of fusing the sub-experimental data according to the similar information to obtain the initial staining parameters corresponding to the data to be stained:

[0138] Determine the similarity between the data to be stained and the sub-experiment data according to the similarity information;

[0139] Calculating the probability value corresponding to the data to be stained belonging to the sub-experiment data according to the similarity;

[0140] Obtaining predicted staining parameters corresponding to staining prediction of the data to be stained under the sub-experimental data;

[0141] The predicted staining parameters are weightedly summed according to the probability value to obtain the initial staining parameters corresponding to the data to be stained.

[0142] In some embodiments, the parameter adjustment module 206 performs, during the process of adjusting the initial staining parameters according to the adherent cell information and the clumping cell information to obtain the target staining parameters corresponding to the data to be stained:

[0143] Performing weighted summation on the adherent cell information according to the probability value to obtain target adherent cell information corresponding to the data to be stained;

[0144] Performing weighted summation on the clumping cell information according to the probability value to obtain target clumping information corresponding to the data to be stained;

[0145] The initial staining parameters are adjusted according to the target wall adhesion information and the target clumping information to obtain the target staining parameters corresponding to the data to be stained.

[0146] In some embodiments, the staining prediction system 200 for cell flow cytometry can be applied to a terminal device.

[0147] It should be noted that, those skilled in the art can clearly understand that, for the sake of convenience and brevity of description, the specific working process of the staining prediction system 200 for cell flow cytometry described above can refer to the corresponding process in the aforementioned embodiment of the staining prediction method for cell flow cytometry, and will not be repeated here.

[0148] See also Figure 3 , Figure 3 A schematic block diagram of the structure of a terminal device provided in an embodiment of the present application.

[0149] like Figure 3 As shown, the terminal device 300 includes a processor 301 and a memory 302, and the processor 301 and the memory 302 are connected via a bus 303, such as an I2C (Inter-integrated Circuit) bus.

[0150] Specifically, the processor 301 is used to provide computing and control capabilities to support the operation of the entire terminal device. The processor 301 can be a central processing unit (CPU), and the processor 301 can also be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. Among them, the general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc.

[0151] Specifically, the memory 302 may be a Flash chip, a read-only memory (ROM) disk, an optical disk, a USB flash drive, or a mobile hard disk.

[0152] Those skilled in the art will understand that Figure 3 The structure shown in the figure is only a block diagram of a partial structure related to the embodiment of the present application, and does not constitute a limitation on the terminal device to which the embodiment of the present application is applied. The specific server may include more or fewer components than shown in the figure, or combine certain components, or have a different arrangement of components.

[0153] The processor is used to run a computer program stored in the memory, and implement any one of the cell flow cytometry staining prediction methods provided in the embodiments of the present application when executing the computer program.

[0154] In one embodiment, the processor is used to run a computer program stored in the memory, and implements the following steps when executing the computer program:

[0155] Obtaining historical experimental data, and performing data clustering on the historical experimental data to obtain a plurality of sub-experimental data;

[0156] Obtaining associated image information corresponding to the sub-experimental data from the historical experimental data;

[0157] Performing edge recognition on the associated image information to obtain the adherent cell information and the clumping cell information corresponding to the sub-experimental data;

[0158] Obtaining data to be stained, and calculating similarity information between the data to be stained and the sub-experiment data;

[0159] The sub-experiment data are subjected to information fusion according to the similar information to obtain initial staining parameters corresponding to the data to be stained;

[0160] Adjusting the initial staining parameters according to the adherent cell information and the clumping cell information to obtain target staining parameters corresponding to the data to be stained;

[0161] Performing dyeing prediction on the data to be dyed according to the target dyeing parameters to obtain a dyeing prediction result corresponding to the data to be dyed.

[0162] In some implementations, the historical experimental data includes at least a first experimental object and a second experimental object. In the process of clustering the historical experimental data to obtain a plurality of sub-experimental data, the processor 301 executes:

[0163] Obtaining a first staining result of the first experimental object in a first time series from the historical experimental data and obtaining a second staining result of the second experimental object in a second time series from the historical experimental data;

[0164] Performing data fitting on the first experimental object according to the first time series and the first staining result to obtain a first fitting result corresponding to the first experimental object;

[0165] Performing data fitting on the second experimental object according to the second time series and the second staining result to obtain a second fitting result corresponding to the second experimental object;

[0166] Data clustering is performed on the first experimental object and the second experimental object according to the first fitting result and the second fitting result to obtain a plurality of sub-experimental data.

[0167] In some embodiments, the associated image information includes an initial adherent image and a plane image, and the processor 301, in the process of performing edge recognition on the associated image information to obtain the adherent cell information and clumping cell information corresponding to the sub-experiment data, executes:

[0168] Performing registration and fusion on any two sub-images in the initial wall-attached image to obtain a connection order corresponding to the initial wall-attached image;

[0169] splicing the sub-images according to the connection order to obtain a target wall-attached image;

[0170] Performing edge recognition on the target adherent image to obtain the adherent cell information corresponding to the target adherent image;

[0171] Performing edge recognition on the plane image to obtain an edge image corresponding to the plane image, and performing image segmentation on the edge image to obtain a segmented object;

[0172] Density calculation is performed according to the segmented object to obtain the agglomerated cell information corresponding to the plane image.

[0173] In some implementations, the processor 301, in the process of stitching the sub-images according to the connection order to obtain the target wall-attached image, executes:

[0174] Obtaining adjacent left and right images from the sub-images according to the connection order;

[0175] According to the registration fusion result, the left image is cut to obtain a first registration area, and the right image is cut to obtain a second registration area;

[0176] Calculating a first gradient corresponding to the first registration area and calculating a second gradient corresponding to the second registration area;

[0177] Determine a corresponding gradient difference value according to the first gradient and the second gradient;

[0178] Performing brightness difference analysis on the first registration area and the second registration area to obtain a brightness difference value;

[0179] fusing the gradient difference value and the brightness difference value to determine a corresponding registration difference value between the first registration area and the second registration area;

[0180] Adjusting the first registration area and the second registration area according to the registration difference to obtain a splicing rule corresponding to the left image and the right image;

[0181] Determine the target wall-attached image according to the splicing rule and the connection order;

[0182] The registration difference is obtained according to the following formula:

[0183]

[0184] Wherein, c represents the registration difference, w1 represents the weight information corresponding to the gradient difference value, w2 represents the weight information corresponding to the brightness difference value, L x1 represents the gradient of the left image in the x direction, L y1 represents the gradient of the left image in the y direction, L x2 represents the gradient of the right image in the x direction, L y2 Represents the gradient of the right image in the y direction, img 1krepresents the brightness information of the left image in k dimension, img 2k represents the brightness information of the right image in the k-dimensional space, n represents the number of dimensional divisions corresponding to the left image or the number of dimensional divisions corresponding to the right image, q k Represents the weight information under the kth dimension.

[0185] In some embodiments, during the process of performing density calculation according to the segmented object to obtain the agglomerated cell information corresponding to the plane image, the processor 301 executes:

[0186] Determine a split window, slide the split window in the plane image, and obtain a window image;

[0187] Obtaining an associated object corresponding to the window image from the segmented image according to the plane image;

[0188] Performing density calculation according to the associated object to obtain density information corresponding to the window image;

[0189] When the density information is greater than or equal to a preset value, performing area calculation on the associated object in the window image to obtain an area value;

[0190] The associated objects are screened out according to the area values ​​to obtain the clumping cell information.

[0191] In some embodiments, the processor 301 performs, during the process of fusing the sub-experimental data according to the similar information to obtain the initial staining parameters corresponding to the data to be stained:

[0192] Determine the similarity between the data to be stained and the sub-experiment data according to the similarity information;

[0193] Calculating the probability value corresponding to the data to be stained belonging to the sub-experiment data according to the similarity;

[0194] Obtaining predicted staining parameters corresponding to staining prediction of the data to be stained under the sub-experimental data;

[0195] The predicted staining parameters are weightedly summed according to the probability value to obtain the initial staining parameters corresponding to the data to be stained.

[0196] In some embodiments, during the process of adjusting the initial staining parameters according to the adherent cell information and the clumping cell information to obtain the target staining parameters corresponding to the data to be stained, the processor 301 executes:

[0197] Performing weighted summation on the adherent cell information according to the probability value to obtain target adherent cell information corresponding to the data to be stained;

[0198] Performing weighted summation on the clumping cell information according to the probability value to obtain target clumping information corresponding to the data to be stained;

[0199] The initial staining parameters are adjusted according to the target wall adhesion information and the target clumping information to obtain the target staining parameters corresponding to the data to be stained.

[0200] It should be noted that those skilled in the art can clearly understand that, for the sake of convenience and brevity of description, the specific working process of the terminal device described above can refer to the corresponding process in the aforementioned cell flow cytometry staining prediction method embodiment, and will not be repeated here.

[0201] The embodiments of the present application also provide a storage medium for computer-readable storage, wherein the storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement the steps of any cell flow cytometry staining prediction method provided in the description of the embodiments of the present application.

[0202] The storage medium may be an internal storage unit of the terminal device in the aforementioned embodiment, such as a hard disk or memory of the terminal device. The storage medium may also be an external storage device of the terminal device, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. equipped on the terminal device.

[0203] It will be appreciated by those skilled in the art that all or some of the steps, systems, and functional modules / units in the methods disclosed above may be implemented as software, firmware, hardware, and appropriate combinations thereof. In a hardware embodiment, the division between the functional modules / units mentioned in the above description does not necessarily correspond to the division of physical components; for example, a physical component may have multiple functions, or a function or step may be performed by several physical components in cooperation. Some or all physical components may be implemented as software executed by a processor, such as a central processing unit, a digital signal processor, or a microprocessor, or implemented as hardware, or implemented as an integrated circuit, such as an application-specific integrated circuit. Such software may be distributed on a computer-readable medium, which may include a computer storage medium (or non-transitory medium) and a communication medium (or temporary medium). As known to those skilled in the art, the term computer storage medium includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules, or other data). Computer storage media include, but are not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital versatile disks (DVD) or other optical disk storage, magnetic cassettes, magnetic tapes, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to store the desired information and can be accessed by a computer. In addition, it is well known to those of ordinary skill in the art that communication media typically contain computer-readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave or other transport mechanism, and may include any information delivery media.

[0204] It should be understood that the term "and / or" used in the present specification and the appended claims refers to any combination of one or more of the associated listed items and all possible combinations, including these combinations. It should be noted that, in this article, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or system including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method, article or system. In the absence of further restrictions, an element defined by the sentence "including a..." does not exclude the presence of other identical elements in the process, method, article or system including the element.

[0205] The serial numbers of the embodiments of the present application are only for description and do not represent the advantages and disadvantages of the embodiments. The above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed by the present invention, and these modifications or replacements should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be based on the protection scope of the claims.

Claims

1. A staining prediction method for cell flow cytometry, characterized in that: The method comprises: Obtaining historical experimental data, and performing data clustering on the historical experimental data to obtain a plurality of sub-experimental data; Obtaining associated image information corresponding to the sub-experimental data from the historical experimental data; Performing edge recognition on the associated image information to obtain the adherent cell information and the clumping cell information corresponding to the sub-experimental data; Obtaining data to be stained, and calculating similarity information between the data to be stained and the sub-experiment data; The sub-experiment data are subjected to information fusion according to the similar information to obtain initial staining parameters corresponding to the data to be stained; Adjusting the initial staining parameters according to the adherent cell information and the clumping cell information to obtain target staining parameters corresponding to the data to be stained; Performing dyeing prediction on the data to be dyed according to the target dyeing parameters to obtain a dyeing prediction result corresponding to the data to be dyed.

2. The method according to claim 1, characterized in that The historical experimental data at least includes a first experimental object and a second experimental object, and the data clustering is performed on the historical experimental data to obtain a plurality of sub-experimental data, including: Obtaining a first staining result of the first experimental object in a first time series from the historical experimental data and obtaining a second staining result of the second experimental object in a second time series from the historical experimental data; Performing data fitting on the first experimental object according to the first time series and the first staining result to obtain a first fitting result corresponding to the first experimental object; Performing data fitting on the second experimental object according to the second time series and the second staining result to obtain a second fitting result corresponding to the second experimental object; Data clustering is performed on the first experimental object and the second experimental object according to the first fitting result and the second fitting result to obtain a plurality of sub-experimental data.

3. The method according to claim 1, characterized in that The associated image information includes an initial adherent image and a plane image, and the step of performing edge recognition on the associated image information to obtain the adherent cell information and clumping cell information corresponding to the sub-experimental data includes: Performing registration and fusion on any two sub-images in the initial wall-attached image to obtain a connection order corresponding to the initial wall-attached image; splicing the sub-images according to the connection order to obtain a target wall-attached image; Performing edge recognition on the target adherent image to obtain the adherent cell information corresponding to the target adherent image; Performing edge recognition on the plane image to obtain an edge image corresponding to the plane image, and performing image segmentation on the edge image to obtain a segmented object; Density calculation is performed according to the segmented object to obtain the agglomerated cell information corresponding to the plane image.

4. The method according to claim 3, characterized in that The step of splicing the sub-images according to the connection sequence to obtain a target wall-attached image includes: Obtaining adjacent left and right images from the sub-images according to the connection order; According to the registration fusion result, the left image is cut to obtain a first registration area, and the right image is cut to obtain a second registration area; Calculating a first gradient corresponding to the first registration area and calculating a second gradient corresponding to the second registration area; Determine a corresponding gradient difference value according to the first gradient and the second gradient; Performing brightness difference analysis on the first registration area and the second registration area to obtain a brightness difference value; fusing the gradient difference value and the brightness difference value to determine a corresponding registration difference value between the first registration area and the second registration area; Adjusting the first registration area and the second registration area according to the registration difference to obtain a splicing rule corresponding to the left image and the right image; Determine the target wall-attached image according to the splicing rule and the connection order; The registration difference is obtained according to the following formula: Wherein, c represents the registration difference, w1 represents the weight information corresponding to the gradient difference value, w2 represents the weight information corresponding to the brightness difference value, L x1 represents the gradient of the left image in the x direction, L y1 represents the gradient of the left image in the y direction, L x2 represents the gradient of the right image in the x direction, L y2 Represents the gradient of the right image in the y direction, img 1k represents the brightness information of the left image in k dimension, img 2k represents the brightness information of the right image in the k-dimensional space, n represents the number of dimensional divisions corresponding to the left image or the number of dimensional divisions corresponding to the right image, q k Represents the weight information under the kth dimension.

5. The method according to claim 3, characterized in that: The step of performing density calculation according to the segmented object to obtain the agglomerated cell information corresponding to the plane image includes: Determine a split window, slide the split window in the plane image, and obtain a window image; Obtaining an associated object corresponding to the window image from the segmented image according to the plane image; Performing density calculation according to the associated object to obtain density information corresponding to the window image; When the density information is greater than or equal to a preset value, performing area calculation on the associated object in the window image to obtain an area value; The associated objects are screened out according to the area values ​​to obtain the clumping cell information.

6. The method according to claim 1, characterized in that The step of fusing the sub-experimental data according to the similar information to obtain initial staining parameters corresponding to the data to be stained includes: Determine the similarity between the data to be stained and the sub-experiment data according to the similarity information; Calculating the probability value corresponding to the data to be stained belonging to the sub-experiment data according to the similarity; Obtaining predicted staining parameters corresponding to staining prediction of the data to be stained under the sub-experimental data; The predicted staining parameters are weightedly summed according to the probability value to obtain the initial staining parameters corresponding to the data to be stained.

7. The method according to claim 6, characterized in that The step of adjusting the initial staining parameters according to the adherent cell information and the clumping cell information to obtain target staining parameters corresponding to the data to be stained includes: Performing weighted summation on the adherent cell information according to the probability value to obtain target adherent cell information corresponding to the data to be stained; Performing weighted summation on the clumping cell information according to the probability value to obtain target clumping information corresponding to the data to be stained; The initial staining parameters are adjusted according to the target wall adhesion information and the target clumping information to obtain the target staining parameters corresponding to the data to be stained.

8. A staining prediction system for cell flow cytometry, characterized in that: include: A data clustering module is used to obtain historical experimental data and perform data clustering on the historical experimental data to obtain multiple sub-experimental data; An image acquisition module, used to obtain associated image information corresponding to the sub-experiment data from the historical experiment data; An edge processing module, used for performing edge recognition on the associated image information to obtain the adherent cell information and the clumping cell information corresponding to the sub-experiment data; A similarity calculation module, used for obtaining the data to be stained and calculating the similarity information between the data to be stained and the sub-experiment data; A parameter determination module, used for fusing the sub-experiment data according to the similar information to obtain initial staining parameters corresponding to the data to be stained; A parameter adjustment module, used for adjusting the initial staining parameters according to the adherent cell information and the clumping cell information to obtain target staining parameters corresponding to the data to be stained; The dyeing prediction module is used to perform dyeing prediction on the data to be dyed according to the target dyeing parameters to obtain a dyeing prediction result corresponding to the data to be dyed.

9. A terminal device, characterized in that: The terminal device includes a processor and a memory; The memory is used to store computer programs; The processor is used to execute the computer program and implement the staining prediction method for cell flow cytometry detection according to any one of claims 1 to 7 when executing the computer program.

10. A computer storage medium for computer storage, characterized in that: The computer storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement the steps of the staining prediction method for cell flow cytometry detection described in any one of claims 1 to 7.

Images