Intelligent iPS precursor cell capturing and identifying method and device, electronic equipment and storage medium

Abstract

According to the iPS precursor cell intelligent capture and identification method and device, the electronic equipment and the storage medium provided by the invention, the cell segmentation model is trained according to the training set cell image, other data sets are predicted according to the cell segmentation model, the cells are tracked according to the prediction result, and the tracking video is generated according to the tracking result. The method comprises the steps of obtaining a tracking video, extracting cell classification features according to the tracking video, training an iPS precursor prediction classification model according to the cell classification features, and predicting and identifying iPS precursor cells according to the iPS precursor prediction classification model. Compared with other technologies only involving part of intermediate tasks, the method has the advantages that a more comprehensive and coherent solution can be provided, the high efficiency and accuracy of cell analysis are ensured, the requirement of manual intervention is reduced, and the stability and repeatability of the cell recognition and tracking process are improved.

Description

Technical Field

[0001] The present application relates to the technical field of cell image segmentation, and particularly relates to a method, device, electronic device and storage medium for intelligent capture and recognition of iPS precursor cells. Background Art

[0002] iPS cells are cells that are reprogrammed from mouse embryonic fibroblasts (MEF) or human fibroblasts into cells with an embryonic-like state by adding induction factors. With the discovery of human iPS cell technology, patient-derived stem cells have great potential in regenerative medicine. It has become an important method for clinical cell therapy and has revolutionized regenerative medicine. For example, the treatment of diseases such as thrombocytopenia, spinal cord injury, macular degeneration, Parkinson's disease, and Alzheimer's disease all apply this technology. However, reprogramming still poses technical and biomedical challenges, including the risk of teratoma formation, and the proportion of iPS precursor cells is very low in the early stage, less than 5%, so that the number of iPS cells that can be obtained in the experimental determination stage is very small. At the same time, the lack of biomarkers to label these cells makes it very laborious and costly to find iPS cells. Recently, there have been some studies on iPS precursor cells, which rely on the analysis of cell detection and cell tracking results. By analyzing the characteristics of cell movement trajectories, it is possible to identify iPS precursor cells in the early reprogramming stage, thus facilitating experimentalists to better understand the characteristics of iPS cells.

[0003] Methods commonly used to study and identify cell images first require cell detection, segmentation, and tracking. Magnusson et al. proposed a global trajectory linking algorithm that links the trajectory cell contours generated by the segmentation algorithm into it. It is a batch processing algorithm that uses the entire image sequence to determine the linking. Grah et al. proposed MitosisAnalyser, a framework based on mathematical imaging methods for detecting, classifying, and tracking mitotic cells in live cell phase contrast imaging. Rea et al. proposed an algorithm based on a graphics processing unit (GPU) for real-time tracking of yeast cells in PhC microscope images to identify and capture yeast cells. The method of tracking through detection determines the minimum cost configuration given by the solution of a linear programming (LP) problem for each pair of frames. Tsai et al. introduced Usiigaci, a semi-automated pipeline for segmenting, tracking, and visualizing cells in PhC sequences. The segmentation is based on mask R-CNN, and the tracking module relies on the Trackpy library. Zhang et al. used traditional machine learning methods to identify precursor cells with a minimum accuracy rate higher than 52%, and also confirmed the differences in the morphology and motion patterns between precursor cells and ordinary cells. The method can provide new technologies or new attempts for improving the efficiency of iPS reprogramming and studying the potential mechanisms of iPS reprogramming.

[0004] There are many deficiencies in the prior art in the field of cell analysis. In terms of feature extraction, it heavily relies on manually designed cell morphological features and cannot effectively capture the complex dynamic characteristics of iPS precursor cells, resulting in low recognition accuracy. At the same time, some cell feature extraction algorithms are limited by closed-source software and cannot be truly integrated into a specific system. In cell division detection, the tracking algorithm has obvious defects. The batch processing mode cannot detect division in real time, and traditional tools lack a modeling mechanism, resulting in a high missed detection rate. In the utilization of spatio-temporal features, the temporal correlation of cell behavior is generally ignored, and most are based on single-frame optimization or only focus on the moment of division, making it difficult to comprehensively model. In terms of system integration, each link is independent and lacks integration, resulting in a large amount of manual intervention and data conversion in the experimental process, which is out of touch with clinical applications. Summary of the Invention

[0005] In view of this, the present invention provides a method, device, electronic device, and storage medium for intelligent capture and recognition of iPS precursor cells, which can ensure the efficiency and accuracy of cell analysis, reduce the need for manual intervention, and improve the stability and repeatability of the cell recognition and tracking process.

[0006] To solve the above problems, the present application adopts the following technical solutions:

[0007] One of the purposes of the present application is to provide a method for intelligent capture and recognition of iPS precursor cells, and the method includes the following steps:

[0008] Obtain training set cell images;

[0009] Train a cell segmentation model based on the training set cell images;

[0010] Perform predictions on other data sets according to the cell segmentation model;

[0011] Track cells according to the prediction results;

[0012] Generate a tracking video according to the tracking results;

[0013] Extract cell classification features according to the tracking video;

[0014] Train an iPS precursor prediction classification model according to the cell classification features;

[0015] Predict and identify iPS precursor cells according to the iPS precursor prediction classification model.

[0016] In some embodiments, in the step of obtaining training set cell images, the following steps are specifically included:

[0017] Obtain images of iPS cells in the early reprogramming stage;

[0018] Perform filtering and denoising processing on the obtained images;

[0019] Perform data augmentation processing on the filtered and denoised images to obtain training set cell images.

[0020] In some embodiments, in the step of training a cell segmentation model based on the training set cell images, the cell segmentation model fuses a Transformer module and a U-Net architecture.

[0021] In some embodiments, in the step of performing predictions on other data sets according to the cell segmentation model, the following steps are specifically included: Process other cell data to make its format consistent with the training set cell images, and input it into the cell segmentation model to automatically generate corresponding segmentation prediction results.

[0022] In some embodiments, in the step of tracking cells according to the prediction results, the following steps are specifically included: Use cross-sequence frame tracking technology to track cells and use a graph neural network to model the temporal association between cells to identify characteristic changes during cell division.

[0023] In some embodiments, in the step of generating a tracking video according to the tracking results, the following steps are specifically included: Draw a border for each cell to mark its position and add annotations to iPS precursor cells, and at the same time visually generate a continuous-frame cell tracking video.

[0024] In some of these embodiments, in the step of extracting cell classification features from the tracking video, the following steps are specifically included: respectively extract the physical features of iPS precursor cells and ordinary cells, and the physical features cover cell information in multiple dimensions, including the speed, volume, sphericity, nuclear-cytoplasmic volume ratio, displacement, area, and standard deviation of intensity of the cells

[0025] In some of these embodiments, in the step of training an iPS precursor prediction classification model based on the cell classification features, the following steps are specifically included:

[0026] Construct a feature data set according to the physical features; use a machine learning algorithm to train the feature data set to distinguish the iPS precursor cells from other types of cells.

[0027] In some of these embodiments, in the step of predicting and identifying iPS precursor cells based on the iPS precursor prediction classification model, the following steps are specifically included: adopt the iPS precursor prediction classification model to identify individuals that conform to the characteristics of iPS precursor cells by analyzing the physical features of each cell.

[0028] The second object of this application is to provide an intelligent capture and identification device for iPS precursor cells, and the device includes:

[0029] An image acquisition unit for acquiring training set cell images;

[0030] A first training unit for training a cell segmentation model according to the training set cell images;

[0031] A first prediction unit for predicting other data sets according to the cell segmentation model;

[0032] A tracking unit for tracking cells according to the prediction results;

[0033] A video unit for generating a tracking video according to the tracking results;

[0034] A feature extraction unit for extracting cell classification features from the tracking video;

[0035] A second training unit for training an iPS precursor prediction classification model according to the cell classification features;

[0036] A second prediction unit for predicting and identifying iPS precursor cells according to the iPS precursor prediction classification model.

[0037] The third object of this application is to provide an electronic device, including at least one processor, at least one memory, and at least one communication bus. Among them, a computer program is stored on the memory, and the processor reads the computer program in the memory through the communication bus;

[0038] When the computer program is executed by the processor, it implements the intelligent capture and recognition method of iPS precursor cells described above.

[0039] A fourth object of the present application is to provide a storage medium on which a computer program is stored, characterized in that when the computer program is executed by a processor, it implements the intelligent capture and recognition method of iPS precursor cells.

[0040] The present application adopts the above technical solutions, and the beneficial effects are as follows:

[0041] The intelligent capture and recognition method, device, electronic device and storage medium of iPS precursor cells provided by the present application train a cell segmentation model according to the training set of cell images, predict other data sets according to the cell segmentation model, track cells according to the prediction results, generate a tracking video according to the tracking results, extract cell classification features according to the tracking video, train an iPS precursor prediction classification model according to the cell classification features, and predict and identify iPS precursor cells according to the iPS precursor prediction classification model. The present application covers the whole process from cell image preprocessing, cell segmentation, cell tracking to precursor cell classification prediction. Compared with other technologies that only involve some intermediate tasks, it can provide a more comprehensive and coherent solution, ensure the efficiency and accuracy of cell analysis, reduce the need for manual intervention, and improve the stability and repeatability of the cell recognition and tracking process.

[0042] In addition, the intelligent capture and recognition method and device of iPS precursor cells provided by the present application use a cell segmentation framework based on the combination of Transformer and Unet, and can accurately extract cell boundaries in high-resolution microscope images. Especially in the case of complex backgrounds and cell adhesions, it shows better segmentation effects than existing methods.

[0043] In addition, the intelligent capture and recognition method and device of iPS precursor cells provided by the present application can effectively maintain the consistency of cells in different time frames through cross-sequence frame tracking technology and GNN model. Especially during cell division, it can accurately track the changes of cells. This has significant advantages over traditional methods in cell tracking tasks, especially in terms of tracking accuracy and stability in long-term sequence videos. Description of the Drawings

[0044] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings required for the description of the embodiments of the present application or the prior art. Obviously, the following described drawings are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained according to these drawings.

[0045] Figure 1 This is the flowchart of the steps of the iPS precursor cell intelligent capture and recognition method provided in Embodiment 1 of the present application.

[0046] Figure 2 This is the structural diagram of the cell segmentation model provided in Embodiment 1 of the present application.

[0047] Figure 3 This is the structural schematic diagram of the iPS precursor cell intelligent capture and recognition method provided in Embodiment 2 of the present application.

[0048] Figure 4 This is the structural schematic diagram of the electronic device provided in Embodiment 3 of the present application. Detailed implementation manners

[0049] The embodiments of the present application will be described in detail below. The examples of the embodiments are shown in the drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the drawings are exemplary and are intended to explain the present application, and should not be construed as limiting the present application.

[0050] In the description of the present application, it should be understood that the orientation or positional relationship indicated by the terms "upper", "lower", "horizontal", "inner", "outer", etc. is based on the orientation or positional relationship shown in the drawings, and is only for the convenience of describing the present application and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be construed as limiting the present application.

[0051] In addition, the terms "first" and "second" are only used for descriptive purposes, and should not be construed as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include one or more of such features. In the description of the present application, "a plurality" means two or more unless otherwise specifically defined.

[0052] In order to make the purpose, technical solutions and advantages of the present application more clear and understandable, the present application will be further described in detail below with reference to the drawings and embodiments.

[0053] Embodiment 1

[0054] Please refer to Figure 1 , which is the flowchart of the steps of the iPS precursor cell intelligent capture and recognition method provided in Embodiment 1 of the present application, including the following steps S110 to step S160. The technical solutions implemented are described in detail below.

[0055] Step S110: Obtain training set cell images.

[0056] In this embodiment, in the step of obtaining the training set cell images, the following steps are specifically included:

[0057] Step S111: Obtain images of iPS cells in the early reprogramming stage.

[0058] In this embodiment, the dataset of the training set cell images is the images taken of iPS cells in the early reprogramming stage. The taken images include bright-field images and red fluorescence images. Among them, the dataset contains a total of 33 fields of view. The first three stages of each field of view contain 233 bright-field images and red fluorescence images.

[0059] Step S112: Perform filtering and denoising processing on the obtained images.

[0060] It can be understood that the background of the cell bright-field images provided in this embodiment has complex textures similar to those of the cytoplasm. There are also pixel value changes in the image background. At the same time, severe cell adhesion makes it impossible to accurately define the boundaries of the cells. These characteristics of the bright-field dataset all bring difficulties to the detection and segmentation of cells. The red fluorescence images are taken under the imaging conditions of point light sources, and there is a problem of point light source imaging interference. Conventional binarization methods cannot be normally used for the fluorescence images taken under point light sources, which brings difficulties to the processing work. To solve the problems existing in the above cell images, the tif function is used to read and parse the files, then filtering is performed, and then noise is removed through connected domain area discrimination to obtain cell images that can be used for subsequent training.

[0061] Furthermore, for the red fluorescence images, CLAHE + bilateral filtering processing is adopted to overcome the problem of point light source imaging interference existing in the red fluorescence images.

[0062] Step S113: Perform data augmentation processing on the filtered and denoised images to obtain the training set cell images.

[0063] In this embodiment, a series of methods can be adopted, such as central cropping, random rotation, elastic transformation, grid distortion, brightness and contrast adjustment, vertical flipping, horizontal flipping, and contrast-limited adaptive histogram equalization (CLAHE), etc. The training dataset is expanded through these methods. After this series of data augmentation operations, the data volume of the training set reaches 14 times the original.

[0064] It can be understood that through data augmentation processing, the robustness of the model can be enhanced, enabling the model to have more stable and reliable performance when facing various different cell image situations.

[0065] Step S120: Train a cell segmentation model according to the training set cell images.

[0066] In this embodiment, the cell segmentation model integrates the Transformer module and the U-Net architecture.

[0067] Please refer to Figure 2 , which is the structural diagram of the cell segmentation model provided in this embodiment. This embodiment adopts a segmentation framework based on Transformer, draws on the design ideas of Swin-Unet and ViT, combines the U-Net structure with parameter reduction and the convolutional residual connection mechanism, alleviates the gradient problem through cross-layer fusion, and realizes the accurate extraction of cell boundaries in high-resolution microscope images; both the encoder and the decoder are embedded with the Transformer module, using the multi-head self-attention mechanism and convolutional transformation to replace the traditional positional encoding, so as to capture global context information and achieve multi-scale feature extraction. At the same time, bilinear interpolation is used for upsampling during the decoding process to ensure the effective recovery of the feature map resolution.

[0068] It should be noted that: The cell segmentation model trained according to the training set cell images provided in this embodiment is implemented based on the Python and PyTorch frameworks and is suitable for cell segmentation tasks of processing large-scale microscope datasets.

[0069] Step S130: Predict other datasets according to the cell segmentation model.

[0070] In this embodiment, in the step of predicting other datasets according to the cell segmentation model, it specifically includes the following steps: processing other cell data to make its format consistent with the training set cell images, and inputting it into the cell segmentation model to automatically generate corresponding segmentation prediction results.

[0071] Specifically, the trained segmentation model is migrated to other datasets. First, the microscopic cell data in other fields of view is preprocessed in the same way to make its format consistent with the training data, and then the data is input into the model, and the model automatically generates corresponding segmentation prediction results. Throughout the process, the model utilizes the learned Transformer and U-Net feature extraction capabilities to accurately identify and label cell boundaries in different data environments, and finally outputs the predicted segmentation results for subsequent analysis and evaluation.

[0072] Step S140: Track cells according to the prediction results.

[0073] In this embodiment, in the step of tracking cells according to the prediction results, it specifically includes the following steps: adopting cross-sequence frame tracking technology to track cells and using graph neural networks to model the temporal associations between cells to identify the characteristic changes during cell division.

[0074] Specifically, first, by analyzing the position changes of cells on the time axis, the identity of each cell is maintained consistent in consecutive frames. Then, cross-sequence frame tracking technology is adopted to track cells, ensuring that the same cells can be consistently labeled in different frames, so as to accurately track cells in consecutive sequence frames.

[0075] Furthermore, in order to detect the process of cell division, a graph neural network (GNN) is introduced to model the temporal correlations between cells through the graph neural network and identify the characteristic changes during cell division.

[0076] Particularly, in the labeling of iPS precursor cells, GNN can accurately track the division and development status of cells, so as to perform accurate classification and labeling in sequence frames, providing reliable data support for subsequent cell analysis and classification tasks.

[0077] It can be understood that through the cross-sequence frame tracking technology and the GNN model in this embodiment, the consistency of cells in different time frames can be effectively maintained. Especially during the cell division process, the changes of cells can be accurately tracked. This has significant advantages over traditional methods in cell tracking tasks, especially in terms of tracking accuracy and stability in long-time sequence videos.

[0078] Step S150: Generate a tracking video according to the tracking results.

[0079] In this embodiment, in the step of generating a tracking video according to the tracking results, the following steps are specifically included: draw a border for each cell to mark its position and add annotations to iPS precursor cells, and at the same time visually generate a cell tracking video of consecutive frames.

[0080] Specifically, according to the tracking results of the previous step, a border can be drawn for each cell to mark its position, and at the same time a cell tracking video of consecutive frames is visually generated. At the same time, in order to capture and identify iPS cells, annotations are added to iPS cells. Finally, all these tracking results and annotations will be synthesized into a video to show the dynamic changes of cells in the entire time series.

[0081] It can be understood that such a cell tracking video not only helps to analyze the movement trajectories of cells, but also provides visual evidence for subsequent cell behavior research.

[0082] Step S160: Extract cell classification features according to the tracking video.

[0083] In this embodiment, in the step of extracting cell classification features according to the tracking video, the following steps are specifically included: extract the physical features of iPS precursor cells and ordinary cells respectively. The physical features cover cell information in multiple dimensions, including the speed, volume, sphericity, nuclear-to-cytoplasmic volume ratio, displacement, area, and standard deviation of intensity of cells.

[0084] It can be understood that according to the tracking and segmentation results, the physical characteristics of iPS precursor cells and normal cells can be respectively extracted, covering cell information in multiple dimensions, including cell velocity, volume, sphericity, nuclear-cytoplasmic volume ratio, displacement, area, and standard deviation of intensity, etc. Through the extraction of these characteristics, not only the morphological characteristics of cells (such as volume and sphericity) are obtained, but also the dynamic behaviors of cells (such as velocity and displacement) and spatial distribution characteristics (such as area and intensity fluctuation) are analyzed. These characteristics provide multi-angle data support for cell classification and behavior analysis, and lay a foundation for the training of the subsequent iPS cell prediction classification model.

[0085] Step S170: Train an iPS precursor prediction classification model according to the cell classification characteristics.

[0086] In this embodiment, in the step of training the iPS precursor prediction classification model according to the cell classification characteristics, the following steps are specifically included:

[0087] Step S171: Construct a feature data set according to the physical characteristics.

[0088] It can be understood that in order to train the iPS precursor cell prediction classification model, a feature data set is first constructed based on the extracted cell physical characteristics (such as volume, velocity, sphericity, etc.).

[0089] Step S172: Use a machine learning algorithm to train the feature data set to distinguish the iPS precursor cells from other types of cells.

[0090] In this embodiment, using a machine learning algorithm includes training these features with a support vector machine, random forest, or deep learning model, with the goal of distinguishing iPS precursor cells from other types of cells.

[0091] It should be noted that during the training process, the parameters of the model can also be optimized through cross-validation to improve the classification accuracy. At the same time, data augmentation techniques are also adopted to ensure that the model has good generalization ability. Finally, through this model, the possible iPS precursor cells in the data set can be predicted and labeled.

[0092] Step S180: Predict and identify iPS precursor cells according to the iPS precursor prediction classification model.

[0093] In this embodiment, in the step of predicting and identifying iPS precursor cells according to the iPS precursor prediction classification model, the following steps are specifically included: Use the iPS precursor prediction classification model to identify individuals that meet the characteristics of iPS precursor cells by analyzing the physical characteristics of each cell.

[0094] It can be understood that when the trained classification model is applied to the iPS cell dataset in other fields of view, the model can identify individuals that conform to the characteristics of iPS precursor cells by analyzing the physical characteristics of each cell (such as velocity, volume, sphericity, etc.). Based on the feature weights learned during the training process, the model can accurately distinguish iPS precursor cells from other types of cells and classify and label each cell. Finally, through this prediction process, we can effectively identify the iPS precursor cells in the dataset and identify and capture iPS cells in advance in the early stage of iPS induction, thereby increasing the proportion of iPS cells in culture.

[0095] The intelligent capture and recognition method for iPS precursor cells provided in this application covers the entire process from cell image preprocessing, cell segmentation, cell tracking to precursor cell classification prediction. Compared with other technologies that only involve some intermediate tasks, it can provide a more comprehensive and coherent solution, ensuring the efficiency and accuracy of cell analysis, reducing the need for manual intervention, and improving the stability and repeatability of the cell recognition and tracking process.

[0096] In addition, the intelligent capture and recognition method for iPS precursor cells provided in this application uses a cell segmentation framework that combines Transformer and Unet, which can accurately extract cell boundaries in high-resolution microscope images. Especially in the case of complex backgrounds and cell adhesions, it shows better segmentation effects than existing methods.

[0097] Furthermore, the intelligent capture and recognition method for iPS precursor cells provided in this application can effectively maintain the consistency of cells in different time frames through cross-sequence frame tracking technology and GNN model. Especially during cell division, it can accurately track the changes of cells. This has significant advantages over traditional methods in cell tracking tasks, especially in terms of tracking accuracy and stability in long-term sequence videos.

[0098] Example 2

[0099] Please refer to Figure 3 , which is the intelligent capture and recognition device for iPS precursor cells provided in this embodiment. The device includes:

[0100] An image acquisition unit 100 for acquiring cell images of the training set.

[0101] A first training unit 120 for training a cell segmentation model according to the cell images of the training set.

[0102] A first prediction unit 130 for making predictions on other datasets according to the cell segmentation model.

[0103] A tracking unit 140 for tracking cells according to the prediction results.

[0104] A video unit 150 for generating a tracking video according to the tracking result.

[0105] A feature extraction unit 160 for extracting cell classification features according to the tracking video.

[0106] A second training unit 170 for training an iPS precursor prediction classification model according to the cell classification features.

[0107] A second prediction unit 180 for predicting and identifying iPS precursor cells according to the iPS precursor prediction classification model.

[0108] For the iPS precursor cell intelligent capture and recognition device provided in Embodiment 2 of this application, the detailed implementation manner can refer to Embodiment 1 and will not be elaborated here.

[0109] The iPS precursor cell intelligent capture and recognition device provided in this application covers the entire process from cell image preprocessing, cell segmentation, cell tracking to precursor cell classification prediction. Compared with other technologies that only involve some intermediate tasks, it can provide a more comprehensive and coherent solution, ensuring the efficiency and accuracy of cell analysis, reducing the need for manual intervention, and improving the stability and repeatability of the cell recognition and tracking process.

[0110] Embodiment 3

[0111] Please refer to Figure 4 , Figure 4 which is a schematic structural diagram of an electronic device provided in an embodiment of this application. The electronic device includes: one or more processors, one or more memories, one or more communication interfaces, and one or more programs; the one or more programs are stored in the memory and are configured to be executed by the one or more processors.

[0112] The above programs include instructions for performing the following steps:

[0113] Obtain training set cell images;

[0114] Train a cell segmentation model according to the training set cell images;

[0115] Make predictions on other data sets according to the cell segmentation model;

[0116] Track cells according to the prediction results;

[0117] Generate a tracking video according to the tracking results;

[0118] Extract cell classification features according to the tracking video;

[0119] Train an iPS precursor prediction and classification model according to the cell classification features;

[0120] Predict and identify iPS precursor cells according to the iPS precursor prediction and classification model.

[0121] All relevant content of each scenario involved in the above method embodiments can be cited in the function descriptions of the corresponding functional modules, and will not be elaborated here.

[0122] It should be understood that the above-mentioned memory may include a read-only memory and a random access memory, and provide instructions and data to the processor. A part of the memory may also include a non-volatile random access memory. For example, the memory may also store information about the device type.

[0123] In the embodiments of the present application, the processor of the above device may be a central processing unit (CPU), and the processor may also be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0124] It should be understood that the "at least one" involved in the embodiments of the present application refers to one or more, and "a plurality" refers to two or more. "And / or" describes the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B may represent: A exists alone, A and B exist simultaneously, and B exists alone, where A and B may be singular or plural. The character " / " generally represents an "or" relationship between the associated objects before and after. "At least one (item)" or its similar expression below refers to any combination of these items, including any combination of single item (item) or plural items (items). For example, at least one (item) of a, b, or c may represent: a, b, c, a - b, a - c, b - c, or a - b - c, where a, b, and c may be single or multiple.

[0125] In addition, unless otherwise stated, the ordinal numbers such as "first" and "second" mentioned in the embodiments of the present application are used to distinguish multiple objects, and are not used to limit the order, timing, priority, or importance of multiple objects. For example, the first information and the second information are only used to distinguish different information, rather than indicating differences in the content, priority, sending order, or importance of these two types of information.

[0126] In the implementation process, each step of the above method can be completed by the integrated logic circuit of the hardware in the processor or the instructions in the form of software. The steps of the method disclosed in combination with the embodiments of the present application can be directly embodied as being executed and completed by the hardware processor, or executed and completed by the combination of the hardware and software units in the processor. The software unit can be located in a mature storage medium in the art such as random access memory, flash memory, read-only memory, programmable read-only memory, or electrically erasable programmable memory, registers, etc. This storage medium is located in the memory, and the processor executes the instructions in the memory and combines its hardware to complete the steps of the above method. To avoid repetition, it will not be described in detail here.

[0127] The embodiments of the present application also provide a computer storage medium, wherein the computer storage medium stores a computer program for electronic data exchange, and the computer program enables the computer to execute some or all of the steps of any one of the methods described in the above method embodiments.

[0128] The embodiments of the present application also provide a computer program product. The above computer program product includes a non-transitory computer-readable storage medium storing a computer program, and the above computer program is operable to enable the computer to execute some or all of the steps of any one of the methods described in the above method embodiments. The computer program product can be a software installation package.

[0129] It should be noted that, for the foregoing method embodiments, for the sake of simple description, they are all expressed as a series of action combinations. However, those skilled in the art should know that the present application is not limited by the described action sequence, because according to the present application, certain steps can be performed in other sequences or simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily essential to the present application.

[0130] In the above embodiments, the descriptions of the various embodiments have their own emphases. For the parts not detailed in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0131] In several embodiments provided by the present application, it should be understood that the disclosed device can be implemented in other ways. For example, the device embodiments described above are only illustrative. For example, the above division of units is only a logical function division. In actual implementation, there can be other division methods. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed coupling or direct coupling or communication connection to each other can be through some interfaces, and the indirect coupling or communication connection of the device or unit can be in an electrical or other form.

[0132] The units described above as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the embodiments of the present application.

[0133] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0134] If the above-mentioned integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable memory. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art or all or part of the technical solution, can be embodied in the form of a software product, which is stored in a memory and includes several instructions for a computer device (which can be a personal computer, server or TRP, etc.) to execute all or part of the steps of the various embodiments of the present application. The aforementioned memory includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, disk or optical disk and other media that can store program codes.

[0135] A person skilled in the art may understand that all or part of the steps in the various methods of the above embodiments may be completed by instructing the relevant hardware through a program, and the program may be stored in a computer-readable memory, which may include a flash drive, ROM, RAM, a magnetic disk or an optical disk, etc.

[0136] The above are only preferred embodiments of the present application, and only specifically describe the technical principles of the present application. These descriptions are only for explaining the principles of the present application and cannot be interpreted as limiting the scope of protection of the present application in any way. Based on the explanation here, any modifications, equivalent substitutions and improvements made within the spirit and principles of the present application, and other specific implementation methods of the present application that can be associated with the technicians in this field without creative work, should be included in the scope of protection of the present application.

Claims

1. A method for intelligent capture and recognition of iPS precursor cells, characterized in that, The method includes the following steps: Obtain training set cell images; Train a cell segmentation model based on the training set cell images; Predict other data sets according to the cell segmentation model; Track cells according to the prediction results; Generate a tracking video according to the tracking results; Extract cell classification features according to the tracking video; Train an iPS precursor prediction classification model according to the cell classification features; Predict and identify iPS precursor cells according to the iPS precursor prediction classification model.

2. The intelligent capture and recognition method of iPS precursor cells according to claim 1, wherein, In the step of obtaining training set cell images, it specifically includes the following steps: Obtain images of iPS cells in the early reprogramming stage; Perform filtering and denoising processing on the obtained images; Perform data augmentation processing on the filtered and denoised images to obtain training set cell images.

3. The iPS precursor cell intelligent capture and recognition method according to claim 1, characterized in that In the step of training a cell segmentation model according to the training set cell images, the cell segmentation model fuses the Transformer module and the U-Net architecture.

4. The method for intelligent capture and recognition of iPS precursor cells according to claim 1, characterized in that, In the step of predicting other data sets according to the cell segmentation model, it specifically includes the following steps: Process other cell data to make its format consistent with the training set cell images, and input it into the cell segmentation model to automatically generate corresponding segmentation prediction results.

5. The iPS precursor cell intelligent capture and recognition method according to claim 1, wherein In the step of tracking cells according to the prediction results, it specifically includes the following steps: Use cross-sequence frame tracking technology to track cells and use graph neural networks to model the temporal correlations between cells to identify characteristic changes during cell division.

6. The method for intelligent capture and recognition of iPS precursor cells according to claim 5, characterized in that, In the step of generating a tracking video according to the tracking results, it specifically includes the following steps: Draw a border for each cell to mark its position and add annotations to iPS precursor cells, and at the same time visually generate a continuous frame cell tracking video.

7. The iPS precursor cell intelligent capture and recognition method according to claim 6, wherein In the step of extracting cell classification features according to the tracking video, it specifically includes the following steps: Extract the physical features of iPS precursor cells and normal cells respectively. The physical features cover cell information in multiple dimensions, including the speed, volume, sphericity, nuclear-cytoplasmic volume ratio, displacement, area, and intensity standard deviation of the cells.

8. The iPS precursor cell intelligent capture and recognition method according to claim 7, characterized in that, In the step of training an iPS precursor prediction classification model according to the cell classification features, it specifically includes the following steps: Construct a feature data set according to the physical features; Use machine learning algorithms to train the feature data set to distinguish iPS precursor cells from other types of cells.

9. The iPS precursor cell intelligent capture and recognition method according to claim 8, wherein In the step of predicting and identifying iPS precursor cells according to the iPS precursor prediction classification model, it specifically includes the following steps: Use the iPS precursor prediction classification model to identify individuals that meet the characteristics of iPS precursor cells by analyzing the physical features of each cell.

10. An intelligent capture and recognition device for iPS precursor cells, characterized in that, The device includes: An image acquisition unit for obtaining training set cell images; A first training unit for training a cell segmentation model according to the training set cell images; A first prediction unit for predicting other data sets according to the cell segmentation model; A tracking unit for tracking cells according to the prediction results; A video unit for generating a tracking video according to the tracking results; A feature extraction unit for extracting cell classification features according to the tracking video; A second training unit for training an iPS precursor prediction classification model according to the cell classification features; A second prediction unit for predicting and identifying iPS precursor cells according to the iPS precursor prediction classification model.

11. An electronic device, characterized in that, Comprising at least one processor, at least one memory, and at least one communication bus, wherein a computer program is stored on the memory, and the processor reads the computer program in the memory through the communication bus; When the computer program is executed by the processor, the iPS precursor cell intelligent capture and identification method according to any one of claims 1 to 9 is implemented.

12. A storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, the iPS precursor cell intelligent capture and identification method according to any one of claims 1 to 9 is implemented.

Images