CAR-T cell culture monitoring system based on image recognition

Abstract

The application relates to the technical field of medical image analysis, in particular to a CAR-T cell culture monitoring system based on image recognition, which comprises a cell region division module, a cell morphology analysis module, a cell behavior dynamic analysis module and a cell population cooperative monitoring module.In the application, cell image gradient intensity analysis is carried out, cell edge detection is refined, and a segmentation threshold is dynamically adjusted, so that cell boundary extraction is ensured to be high in precision, noise interference is avoided, cell shape, texture and geometric features are refined, the recognition degree of morphological features is improved, cell morphology micro changes are accurately captured, cell behavior is dynamically tracked and trend evaluation is carried out based on adjacent frame images, and the dynamic changes such as cell division, aggregation and migration are accurately reflected, over-smoothing in a dynamic process is avoided, cell population behavior monitoring is combined with interaction and corresponding speed calculation, population cooperative effect change trends are comprehensively evaluated, and more refined cell culture environment optimization and clinical research support are provided.

Description

Technical Field

[0001] The present invention relates to the technical field of medical image analysis, and in particular to a CAR-T cell culture monitoring system based on image recognition. Background Art

[0002] The technical field of medical image analysis includes technical methods based on image processing and analysis and their application in medical scenarios. The core content of this technical field is to assist clinical diagnosis, treatment and research through the acquisition, processing, analysis and interpretation of medical images. Specific technologies include image segmentation, feature extraction, target recognition and classification, etc., which are widely used in pathological analysis, surgical navigation, image evaluation and other fields. The overall technical field of medical image analysis combines computer vision, biomedical engineering and artificial intelligence technologies to process various types of medical image data in a precise and efficient manner, such as microscope images, CT images, MRI images and ultrasound images, providing important technical support for medical scenarios.

[0003] Among them, the CAR-T cell culture monitoring system refers to a system used to monitor and evaluate the culture status in real time during the CAR-T cell culture process through image analysis technology. The system focuses on technical matters such as cell proliferation, differentiation and state changes in CAR-T cell culture. It uses a high-resolution image acquisition device to obtain image data of cells during the culture process, and uses image processing methods to complete cell contour detection, cell number counting and morphological analysis. The system uses image recognition-based methods to analyze and record the specific characteristics of cells during the culture process, so as to realize the determination of cell growth status in the culture environment and the characteristic identification of key stages.

[0004] In the process of image acquisition and analysis, the existing technology relies on traditional cell segmentation methods, which have great limitations in the detailed analysis of cell morphology and behavior. Due to the noise interference of image data and the challenges of cell morphology complexity, the existing technology has difficulty in accurately dividing cell boundaries, resulting in insufficiently detailed extraction of cell morphological features, especially in high-density culture or when cell morphology is highly homogeneous, which easily leads to recognition errors. The existing technology has limited ability to dynamically track cell behavior and relies on image comparison at long time intervals, which results in large errors in capturing subtle changes in cell behavior and cannot reflect the precise dynamics of cells in real time. In terms of group synergy analysis, the existing technology simplifies the assessment of cell relationships and cannot effectively measure the complex interactions and corresponding speeds within the cell group, resulting in a relatively rough analysis of group behavior patterns and an inability to fully reflect changes in group synergy effects. The existing technology has certain deficiencies in the precise monitoring and dynamic analysis of cell behavior and the comprehensive assessment of cell group synergy, which affects the efficiency and accuracy of various operations in the cell culture process. Summary of the Invention

[0005] The purpose of the present invention is to solve the shortcomings of the prior art and propose a CAR-T cell culture monitoring system based on image recognition.

[0006] To achieve the above objectives, the present invention adopts the following technical solutions: A CAR-T cell culture monitoring system based on image recognition includes:

[0007] The cell region segmentation module analyzes the gradient intensity of cell culture images and divides the images into high-feature regions and low-feature regions. It then detects cell edges in high-feature regions, extracts edge positions and cell morphological features, refines cell shape and texture, and dynamically adjusts the segmentation threshold to obtain cell boundary segmentation results.

[0008] The cell morphology analysis module analyzes the morphology of cells in high-feature areas based on the cell boundary segmentation results, extracts local texture information of cells, analyzes cell morphology differences through local feature comparison, refines the geometric and structural features of cells, and obtains a cell morphology feature dataset;

[0009] The cell behavior dynamic analysis module tracks and analyzes cell behavior based on the cell morphological feature dataset, monitors cell division, aggregation, and migration, evaluates distance and behavior trends through adjacent frame images, records cell dynamic behavior, analyzes dynamic changes in cell culture, and obtains cell behavior trend analysis results;

[0010] The cell population collaborative monitoring module analyzes the synergistic effects between cell populations based on the cell behavior trend analysis results, monitors the trends of cell aggregation and dispersion, calculates the distance and corresponding speed between cells based on the relationship between cells, evaluates the overall behavior pattern of the cell population, dynamically monitors the synergistic effects of the cell population, and obtains the cell population collaborative behavior monitoring results.

[0011] As a further solution of the present invention, the step of obtaining the edge position and cell morphological characteristics is specifically as follows:

[0012] Based on the cell culture image, the image is converted into a grayscale image, the brightness value of each pixel and the grayscale difference of the surrounding pixels are analyzed, the gradient value of each pixel is calculated, the gradient value is used to divide the image into regions, and the regional features are judged according to the gradient value to obtain the high and low feature region division results;

[0013] Based on the high- and low-feature region division results, regional enhancement processing is performed on the pixels in the high-feature region, the local contrast of each pixel in the high-feature region is analyzed, and weights are assigned according to the difference in surrounding pixel values ​​to obtain enhanced data for the corresponding region;

[0014] Based on the enhanced data of the corresponding area, the edge position of the cell is located by contour, the gradient change of each pixel in the high-feature area is analyzed, the edge pixel point is determined, and the cell morphology is analyzed according to the edge information to obtain the edge position and cell morphological characteristics.

[0015] As a further solution of the present invention, the steps of obtaining the cell boundary segmentation result are specifically as follows:

[0016] Extracting edge information from the cell image based on the edge position and cell morphology characteristics, identifying the position information of each cell boundary, and generating a cell boundary position feature dataset by analyzing the cell morphology characteristics, including curvature and edge smoothness;

[0017] By combining the cell boundary position feature dataset with cell texture information, texture features in the image are extracted, the changing pattern of texture features in the differentiated edge area is analyzed, and the segmentation threshold is dynamically adjusted to obtain a segmentation threshold adjustment coefficient;

[0018] According to the segmentation threshold adjustment coefficient, the optimal matching of cell morphology and texture features is used, and the formula is adopted:

[0019]

[0020] Generate cell boundary segmentation results;

[0021] Among them, T final represents the cell boundary segmentation threshold, T base represents the initial segmentation threshold, ω i Represents the weight coefficient, X i represents the adjustment coefficient of the association between cell morphology and texture features, and n is the number of features.

[0022] As a further solution of the present invention, the step of acquiring local cell texture information is specifically as follows:

[0023] Based on the cell boundary segmentation results, boundary refinement and correction are performed, irregular edges are removed by local smoothing, the boundaries are repaired in combination with cell contour features, breaks and discontinuities are avoided, and morphology optimization is performed according to pixel distribution to obtain the corrected cell boundaries;

[0024] Based on the corrected cell boundaries, geometric feature extraction is performed, by identifying the area and perimeter of each cell, using the boundary points to calculate the aspect ratio and shape factor, analyzing the symmetry and circularity of the cells, and screening the cell shape based on the area and perimeter ratio to obtain cell size and shape characteristics;

[0025] Based on the cell size and shape characteristics, local areas are selected for texture analysis, the pixel value changes in the local cell area are counted, the degree of texture variation and consistency are quantified, the texture distribution pattern is analyzed in combination with the boundary information, the texture difference and consistency between regions are evaluated, and the local texture information of the cell is obtained.

[0026] As a further embodiment of the present invention, the steps for acquiring the cell morphology feature dataset are specifically as follows:

[0027] According to the local texture information of the cells, the cell edges, texture lines and contrast between regions are extracted, and combined with regional division, the image grayscale changes, local contrast, and texture direction parameters are identified, and the geometric differences between local regions are analyzed to obtain cell texture features;

[0028] Based on the cell texture features, the cell morphology differences and geometric features between differentiated regions are quantified and compared with the texture similarity metric between regions using the formula:

[0029]

[0030] Obtain cell morphology difference data;

[0031] Among them, δ represents the local texture difference, T r,j is the texture feature of the right area, T l,j is the texture feature of the left area, N is the number of texture points;

[0032] Based on the cell morphology difference data, the geometric and structural features of the cells are refined through clustering and segmentation methods, and regions with similar geometric features are grouped using distance measurement to obtain a cell morphology feature dataset.

[0033] As a further embodiment of the present invention, the steps for obtaining the cell behavior trend analysis results are specifically as follows:

[0034] Analyze each frame of the cell morphology feature dataset to extract cell position and identify the distance between adjacent frames, use the distance to judge cell division, aggregation and migration behavior, track cell position through time series and combined with image pixel position to obtain cell behavior tracking data;

[0035] Based on the cell behavior tracking data, the dynamic behavior of the cells was evaluated according to the motion trajectory and change trend between adjacent frames, and the displacement distance value of the cells between adjacent frames was calculated using the formula:

[0036]

[0037] Obtain cell movement trend data;

[0038] Among them, d represents the displacement distance of the cell between adjacent frames, x t+1 and y t+1 is the coordinate of the cell in frame t+1, x t and y t is the coordinate of the cell in frame t;

[0039] Based on the cell movement trend data, the cell division, aggregation and migration behaviors are analyzed, the cell behavior trends are evaluated, and whether there is a changing trend is analyzed. The criticality of the cell behavior change is judged by setting a threshold to obtain the cell behavior trend analysis results.

[0040] As a further embodiment of the present invention, the steps of obtaining the distance between cells and the corresponding speed are specifically as follows:

[0041] Based on the cell behavior trend analysis results, analyzing the positional relationship between cells, extracting the coordinate information of each pair of cells, identifying the straight-line distance between cells, extracting the cell coordinates and tracking the position changes of each pair of cells, and obtaining a distance data set between cells;

[0042] Based on the intercellular distance dataset, the intercellular velocity was calculated according to the displacement of each pair of cells in adjacent frames. The intercellular velocity change was analyzed by taking the derivative of the cell coordinate difference, using the formula:

[0043]

[0044] Get the distance between cells and the corresponding speed;

[0045] Among them, v represents the velocity value of each pair of cells, D t+1 and D t is the intercellular distance between the t+1th frame and the tth frame, and Δt is the time interval.

[0046] As a further embodiment of the present invention, the steps for obtaining the monitoring results of the cell population coordinated behavior are specifically as follows:

[0047] Based on the distance and corresponding speed between the cells, the spatial position coordinates and motion state of each cell are captured, the corresponding speed of the cells is monitored, the distance and speed of each cell to the surrounding cells are recorded, and a cell group position and speed data set is generated;

[0048] Based on the cell population position and velocity dataset, the corresponding position and motion state of each cell and its surrounding cells are analyzed, cells with interaction strength greater than a set threshold are screened, and the interaction range of each pair of cells is defined based on distance, velocity, and interaction time, thereby constructing a dataset of cell-to-cell interaction relationships.

[0049] The intercellular interaction relationship dataset is used to analyze the coordinated behavior of the cell population. Through the interaction between cells and the group trend, the dynamic changes of the group behavior are judged to obtain the monitoring results of the cell population coordinated behavior.

[0050] Compared with the prior art, the advantages and positive effects of the present invention are:

[0051] In the present invention, through gradient intensity analysis of cell images, the images are effectively divided into high-feature regions and low-feature regions, thereby refining cell edge detection. Dynamic adjustment of the segmentation threshold ensures high-precision extraction of cell boundaries without being affected by interference such as light and noise. Cell morphology analysis is performed on high-feature regions to refine cell shape, texture, and geometric features, helping to accurately capture subtle morphological changes in cells, thereby providing more in-depth data on cell status. This targeted and refined analysis effectively improves the recognition and accuracy of cell morphological features. Dynamic tracking and trend assessment of cell behavior provide higher spatiotemporal resolution for real-time monitoring of cell division, aggregation, migration, and other aspects, enabling more accurate capture of cell behavior patterns. Evaluation of cell behavior based on adjacent frame images allows the analysis results to accurately reflect subtle changes in cell trends over a short period of time, effectively avoiding situations where oversmoothing of dynamic processes or neglect of local details is the case. Cell population behavior monitoring, combined with cell-cell interactions and corresponding velocity calculations, can comprehensively assess the changing trends of cell population synergy, providing a panoramic analysis of cell population behavior patterns, providing strong data support for more efficient cell culture environment optimization and clinical research. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 is a system flow chart of the present invention;

[0053] Figure 2 Flowchart showing edge position and cell morphology characteristics in the present invention;

[0054] Figure 3 This is a flow chart of the cell boundary segmentation results in the present invention;

[0055] Figure 4 This is a flow chart of local cell texture information in the present invention;

[0056] Figure 5 This is a flow chart of the cell morphology feature dataset in the present invention;

[0057] Figure 6 This is a flow chart of the cell behavior trend analysis results in the present invention;

[0058] Figure 7 A flow chart showing the distance between cells and the corresponding speed in the present invention;

[0059] Figure 8 Flowchart showing the results of monitoring the coordinated behavior of cell populations in the present invention. DETAILED DESCRIPTION

[0060] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0061] In the description of the present invention, it should be understood that the terms "length," "width," "up," "down," "front," "back," "left," "right," "vertical," "horizontal," "top," "bottom," "inside," "outside," and the like, indicating positions or relationships, are based on the positions or relationships shown in the accompanying drawings and are intended only to facilitate the description of the present invention and simplify the description. They do not indicate or imply that the devices or elements referred to must have a specific orientation, be constructed, or operate in a specific orientation. Therefore, they should not be construed as limiting the present invention. Furthermore, in the description of the present invention, "plurality" means two or more, unless otherwise expressly and specifically defined.

[0062] See also Figure 1 , the CAR-T cell culture monitoring system based on image recognition includes:

[0063] The cell region segmentation module analyzes the gradient intensity of cell culture images and divides the images into high-feature regions and low-feature regions. It then detects cell edges in high-feature regions, extracts edge positions and cell morphological features, refines cell shape and texture, and dynamically adjusts the segmentation threshold to obtain cell boundary segmentation results.

[0064] The cell morphology analysis module analyzes the morphology of cells in high-feature areas based on the cell boundary segmentation results, extracts local cell texture information, analyzes cell morphology differences through local feature comparison, refines the geometric and structural features of cells, and obtains a cell morphology feature dataset;

[0065] The cell behavior dynamic analysis module tracks and analyzes cell behavior based on the cell morphological feature dataset, monitors cell division, aggregation, and migration, evaluates distance and behavior trends through adjacent frame images, records cell dynamic behavior, analyzes dynamic changes in cell culture, and obtains cell behavior trend analysis results;

[0066] The cell population collaborative monitoring module is based on the results of cell behavior trend analysis, analyzes the synergistic effects between cell populations, monitors the trends of cell aggregation and dispersion, combines the relationships between cells, calculates the distance and corresponding speed between cells, evaluates the overall behavior pattern of the cell population, dynamically monitors the synergistic effects of the cell population, and obtains the monitoring results of cell population collaborative behavior.

[0067] The cell boundary segmentation results include cell edge position, cell morphological characteristics, and cell segmentation threshold. The cell morphological feature dataset includes cell shape, cell size, cell edge, local texture, cell geometric characteristics, and cell structural characteristics. The cell behavior trend analysis results include cell division behavior, cell aggregation behavior, cell migration behavior, cell behavior trend, and cell dynamic changes. The cell group collaborative behavior monitoring results include cell aggregation trend, cell dispersion trend, cell distance, cell corresponding speed, cell group behavior pattern, and cell synergistic effect.

[0068] See also Figure 2 , the steps for obtaining edge position and cell morphology characteristics are as follows:

[0069] Based on the cell culture image, the image is converted into a grayscale image, the brightness value of each pixel and the grayscale difference of the surrounding pixels are analyzed, the gradient value of each pixel is calculated, the gradient value is used to divide the image into regions, and the regional features are judged according to the gradient value to obtain the high and low feature region division results;

[0070] The cell culture image is converted into a grayscale image for subsequent analysis. The conversion process relies on extracting the brightness value of each pixel in the original image. The corresponding grayscale value is obtained by weighted averaging the pixel values ​​of the red, green, and blue channels. The grayscale difference between the brightness value of each pixel and the grayscale of the surrounding pixels is analyzed to calculate the gradient value. The gradient value represents the rate of change of pixel intensity in the image and can reflect the edge and change area of ​​the image. When calculating the gradient value for each pixel, a preset size window is selected and the gradient value of the pixel is derived by calculating the difference in the brightness value of each pixel in the window. The specific process can be obtained by calculating the difference in the surrounding pixels. It can quantify the changing and static areas in the cell image and then perform regional segmentation based on the calculated gradient value. By setting a gradient threshold, the image is divided into high-feature areas and low-feature areas. The high-feature area reflects strong cell activity, while the low-feature area is relatively stable. In this way, the high-feature and low-feature area division results of the cell image can be obtained. Through regional gradient feature analysis, the cell culture image can be preliminarily classified, laying the foundation for subsequent cell morphology analysis.

[0071] Based on the results of high and low feature area division, regional enhancement processing is performed on the pixels in the high feature area. The local contrast of each pixel in the high feature area is analyzed, and weights are assigned according to the difference in surrounding pixel values ​​to obtain enhanced data for the corresponding area.

[0072] Enhancement is performed on high-feature areas, especially in cell culture images, which reflect areas of cell activity or large changes. To enhance the recognizability of high-feature areas, the local contrast of each pixel is first calculated. Local contrast can be achieved by analyzing the grayscale difference between the pixel and the surrounding pixels. A window is used to analyze the difference between the pixel in the window and its neighborhood to calculate the local contrast. After obtaining the local contrast, ensure that the enhancement effect conforms to the characteristics of the cell image and weight distribution can be performed based on the contrast. For areas with larger contrast, the weight value is higher. In this way, during the enhancement process, the parts with larger changes will be enhanced more. It can highlight the details of the high-feature area without affecting the low-feature area, making the cell morphology and features in the image more distinct, and obtaining enhanced data for the corresponding area.

[0073] Based on the enhanced data of the corresponding area, the edge position of the cell is located by contour, the gradient change of each pixel in the high-feature area is analyzed, the edge pixel point is determined, and the cell morphology is analyzed based on the edge information to obtain the edge position and cell morphological characteristics;

[0074] In high-feature areas, the gradient changes of each pixel are analyzed. Changes in gradient values ​​can often accurately reflect the presence of edges. By analyzing gradient changes, the location of edge pixels can be clearly determined, and the cell edge can be inferred. Through precise edge positioning, the cell outline can be identified, and further cell morphology can be analyzed to determine cell morphological characteristics, such as cell size and shape. The analysis results can help better understand the growth and distribution of cells, providing a more accurate reference for monitoring CAR-T cell culture. Through detailed analysis of edge information, not only can changes in cell morphology be understood, but the interaction between cells and the environment can also be further explored, thereby promoting real-time monitoring and optimization of the cell culture process and determining edge location and cell morphological characteristics.

[0075] See also Figure 3 , the steps to obtain the cell boundary segmentation results are as follows:

[0076] Based on edge position and cell morphology features, edge information in cell images is extracted, and the position information of each cell boundary is identified. By analyzing cell morphology features, including curvature and edge smoothness, a cell boundary position feature dataset is generated.

[0077] First, edge information is extracted from the cell image, and edge detection is performed through image processing technology. Methods such as image gradient or Canny edge detection are used to extract cell boundaries, extract the edge position of each cell, determine the coordinate points of each cell boundary, and combine cell morphological characteristics, such as morphological characteristics, cell aspect ratio, contour smoothness, etc. to further refine the cell boundary analysis. Morphological filtering or contour analysis is used to calculate features such as cell curvature and edge smoothness to form a morphological feature vector. In this way, it can ensure that the shape changes and edge features of the cells are accurately captured, thereby preparing sufficient morphological data for subsequent cell segmentation and generating a cell boundary position feature dataset. This dataset will serve as the key input in subsequent analysis to ensure that the subsequent steps can dynamically adjust the segmentation threshold based on this dataset.

[0078] By combining the cell boundary position feature dataset with cell texture information, we extract the texture features in the image, analyze the changing patterns of texture features in differentiated edge areas, dynamically adjust the segmentation threshold, and obtain the segmentation threshold adjustment coefficient.

[0079] Combined with the cell texture feature information, the cell texture distribution in the image is further analyzed. The texture features of each cell are extracted through texture analysis methods such as gray-level co-occurrence matrix (GLCM). The features include contrast, homogeneity, energy and entropy, which can reflect the texture information of the cell edge area. The texture features are combined with the edge position feature data to analyze the changing rules of the texture features in different edge areas. The image region segmentation method, such as the K-means clustering algorithm, is used to refine the cell area and identify the different texture areas of the cell. Not only can the position of the cell edge be accurately obtained, but also the texture features of different areas inside and outside the cell can be obtained, providing more accurate data support for threshold adjustment. According to the texture analysis results, a dynamic adjustment method based on entropy value is used to update the segmentation threshold to obtain the segmentation threshold adjustment coefficient.

[0080] According to the segmentation threshold adjustment coefficient, the optimal matching of cell morphology and texture features is used, and the formula is used:

[0081]

[0082] Generate cell boundary segmentation results;

[0083] Among them, T final represents the cell boundary segmentation threshold, T base represents the initial segmentation threshold, ω i Represents the weight coefficient, X i represents the adjustment coefficient of the association between cell morphology and texture features, and n is the number of features;

[0084] The benefit of the formula is that by introducing the weight coefficients of morphological features and texture features and combining the threshold adjustment coefficient, the segmentation threshold can more accurately match the complex features in the cell image, thereby improving the accuracy and reliability of the segmentation results; the initial segmentation threshold T base According to the pre-set basic characteristics of the cell image, the adjustment coefficient X of each feature is calculated by comprehensively considering the cell morphology and texture characteristics. i , and set the weight coefficient ω according to its influence i The coefficients are obtained through experimental data or previous model training, and the weight coefficients are set by comparing the influence of different features on the final segmentation results. For example, if the morphological feature has a greater impact on the segmentation result, a higher weight is assigned;

[0085] The specific calculation process is as follows: final =T base ·(1+(ω1·X1+ω2·X2+…+ω n ·X n ));

[0086] Among them, T base is the initial threshold, ω i is the weight coefficient of the i-th feature, X i is the adjustment coefficient of the i-th feature, and n is the number of features;

[0087] The results show that by dynamically adjusting the segmentation threshold, cell boundaries can be accurately identified in the image, avoiding excessive or insufficient segmentation, and thus improving segmentation accuracy. In this case, if the initial segmentation threshold T base = 100, weight coefficients ω1 = 0.7, ω2 = 0.3, and adjustment coefficients X1 = 5, X2 = 3;

[0088] The final calculated segmentation threshold is: T final =100·(1+(0.7·5+0.3·3))=100·(1+3.9)=490

[0089] Therefore, the final segmentation threshold is 490.

[0090] See also Figure 4 , the steps for obtaining local texture information of cells are as follows:

[0091] Based on the cell boundary segmentation results, boundary refinement and correction are performed. Irregular edges are removed through local smoothing. The boundaries are repaired by combining cell contour features to avoid breaks and discontinuities. Morphological optimization is performed based on pixel distribution to obtain the corrected cell boundaries.

[0092] First, the boundaries in the original segmentation results are smoothed, and small sharp protrusions or noise points are removed by local pixel averaging. The local gradient values ​​of the boundary pixels are calculated to mark sharp changes at the edges, and adjustments are made using a gradually decreasing smoothing window. This process ensures the continuity of the cell boundaries by combining the cell contour features. For broken or discontinuous parts, the boundaries are repaired using interpolation technology based on the spatial distribution of surrounding pixels. Interpolation not only takes into account the pixel adjacency relationship, but also combines the morphological characteristics of the cells themselves, such as circles and ellipses, to ensure that the repaired boundaries conform to the natural morphology of the cells. Morphological optimization is performed, and the boundaries are re-evaluated using the pixel clustering method to eliminate irregular edges caused by erroneous segmentation. Finally, the corrected cell boundaries are obtained, which are smoother and more continuous, reducing morphological distortion caused by segmentation errors.

[0093] Based on the corrected cell boundaries, geometric feature extraction is performed. By identifying the area and perimeter of each cell, the aspect ratio and shape factor are calculated using the boundary points, the symmetry and circularity of the cells are analyzed, and the cell shape is screened based on the area and perimeter ratio to obtain the cell size and shape characteristics.

[0094] First, the area and perimeter of each cell are extracted. These two basic geometric features can be obtained by counting the pixels of the boundary points and the length of the boundary path. The area is obtained by counting the number of pixels in the closed area, while the perimeter is obtained by calculating the distance between the boundary points one by one and summing them up. The aspect ratio of the cell is calculated. The aspect ratio is obtained by identifying the bounding box of the cell and calculating its size. In order to accurately reflect the shape characteristics of the cell, the shape factor is also calculated. This factor combines the ratio of area to perimeter to quantitatively describe the circularity and symmetry of the cell and analyze the symmetry and circularity of the cell shape. If the cell shows high symmetry and a shape close to a circle, it indicates that it is growing well. Combined with the ratio of area to perimeter, the cell shape characteristics that meet the conditions are screened out, and the cell size and shape characteristics are obtained to provide a reference for further biological analysis.

[0095] Based on the size and shape characteristics of cells, local areas are selected for texture analysis. The pixel value changes in the local area of ​​the cell are counted, the degree of texture variation and consistency are quantified, and the texture distribution pattern is analyzed in combination with the boundary information. The texture difference and consistency between regions are evaluated to obtain the local texture information of the cell.

[0096] First, based on the size and shape characteristics of the cell, a local area of ​​the cell is selected. The area represents the active part of the cell or the area with special functions. By counting the pixel values ​​in the local area, the range of change of the pixel values ​​and its distribution are calculated, and the degree of texture variation and consistency are quantified. In order to quantify the degree of texture variation, the roughness and uniformity of the texture can be obtained by calculating the gradient difference of adjacent pixel values ​​in the local area. The consistency is evaluated by statistically analyzing the distribution balance of pixel values. Next, combined with boundary information, the distribution pattern of the local texture is analyzed, the texture characteristics of the area near the boundary are observed, and the texture differences between different areas are compared. This texture difference not only helps to understand the structural characteristics inside the cell, but also reveals the functional status or health status of the cell. The local texture information of the cell is obtained through operation to support further cell analysis and evaluation.

[0097] See also Figure 5 ,The steps for obtaining the cell morphology feature dataset are as follows:

[0098] According to the local texture information of cells, the cell edges, texture lines and contrast between regions are extracted. Combined with regional division, the image grayscale changes, local contrast and texture direction parameters are identified, and the geometric differences between local regions are analyzed to obtain the cell texture characteristics.

[0099] First, image segmentation is performed on each local area. The image is divided into multiple regions of interest using image processing algorithms. The texture features in each region, such as local contrast and gray-level co-occurrence matrix (GLCM), are calculated to extract the texture feature value of each region. The regional texture contrast is further used to reflect structural features such as cell edges or cell membranes. The texture features of each region are extracted by comparing the regional grayscale values. By comparing the features with the original image, the cell morphology differences are analyzed, and finally the cell texture features are obtained to further quantify the texture differences of each region.

[0100] Based on the cell texture features, the cell morphology differences and geometric features between differentiated regions are quantified and compared with the texture similarity metric between regions using the formula:

[0101]

[0102] Obtain cell morphology difference data;

[0103] Among them, δ represents the local texture difference, T r,j is the texture feature of the right area, T l,j is the texture feature of the left area, N is the number of texture points;

[0104] The benefit of the formula is that it can quantify cell morphology differences and improve the accuracy of cell morphology feature extraction through texture contrast analysis;

[0105] δ represents the local texture difference, which measures the morphological difference by calculating the texture difference between the right and left regions of the cell. r,j is the texture feature of the right area, T l,j is the texture feature of the left area, N is the number of texture points, and the square difference of each contrasting texture is summed and averaged to obtain the difference of the area. The texture value of the right area is set to T r,j = [5, 8, 7] and the texture value of the left area is T l,j =[4, 9, 6], calculate according to the formula:

[0106]

[0107] The results show that the texture difference between the right and left regions of the cell is 1, indicating that there are certain differences in local texture features between the two, which in turn affects the analysis of cell morphology difference data.

[0108] Based on the cell morphology difference data, the geometric and structural features of the cells are refined through clustering and segmentation methods. The regions with similar geometric features are grouped using distance measurement to obtain the cell morphology feature dataset.

[0109] A clustering algorithm is used to group the data, and regions with similar geometric and structural features are divided into the same group. A clustering method based on Euclidean distance is used. First, the texture difference of each pair of regions is calculated. The difference calculated above is used as sample data. Through a hierarchical clustering algorithm, similar texture regions are gradually classified together, and finally a group of cell regions with obvious characteristics is generated. Further optimization and integration are performed to obtain a cell morphological feature data set. By setting a threshold, significant areas of cell morphological changes are identified, and areas with larger changes are marked to further refine the geometric structural features of the region.

[0110] See also Figure 6 , the specific steps for obtaining the results of cell behavior trend analysis are as follows:

[0111] Analyze each frame of the cell morphology feature dataset to extract cell position and identify the distance between adjacent frames. Use distance to judge cell division, aggregation, and migration behaviors. Track cell positions through time series and combined with image pixel positions to obtain cell behavior tracking data.

[0112] Each frame of the cell morphological feature dataset needs to extract the position of the cell through image processing methods. The coordinates of the cell can be obtained through the pixel position of the image. For each frame of the image, a reference area is first defined to compare the position changes of the cells in the area. The displacement of the cells is calculated frame by frame, and the pixel difference of the image is used to represent the movement of the cells. The identification of cell division, aggregation, and migration behaviors requires calculating the movement speed and direction of the cell in each time unit based on the position difference between adjacent images of each frame. By calculating the displacement of the cell in each frame, the distance between adjacent frames is obtained. The cell behavior trend is further evaluated by displacement calculation to obtain the dynamic behavior of each cell. Taking a cell as an example, assuming that in the first frame, the coordinates of the cell are (100, 200), and in the second frame, the coordinates of the cell are (110, 210), the distance between these two points is calculated to obtain the movement distance of the cell, and the calculation result finally obtains the cell behavior tracking data.

[0113] Based on the cell behavior tracking data, the dynamic behavior of cells is evaluated according to the motion trajectory and change trend between adjacent frames, and the displacement distance of cells between adjacent frames is calculated using the formula:

[0114]

[0115] Obtain cell movement trend data;

[0116] Among them, d represents the displacement distance of the cell between adjacent frames, x t+1 and y t+1 is the coordinate of the cell in frame t+1, x t and y t is the coordinate of the cell in frame t;

[0117] The benefit of the formula is that by calculating the displacement of cells between adjacent frames, the dynamic behavior of cells can be quantified and the movement trajectory of cells can be accurately tracked;

[0118] d represents the displacement of the cell between adjacent frames, x t+1 and y t+1 is the coordinate of the cell in frame t+1, x t and y t is the coordinate of the cell in the t-th frame. When calculating, first take the cell position coordinates in two adjacent frame images, then calculate with the square of the coordinate difference, and then take the square root to get the actual distance between the two frames. In actual application, assuming that the cell position in the first frame is (100, 200) and the cell position in the second frame is (110, 210), then calculate according to the formula:

[0119]

[0120] The results show that the cell movement distance between two adjacent frames is about 14.14 pixels, which further reflects the dynamic change trend of cells and helps to analyze cell division, aggregation, migration and other behaviors.

[0121] Based on cell movement trend data, analyze cell division, aggregation and migration behaviors, evaluate cell behavior trends, analyze whether there is a changing trend, and determine the criticality of cell behavior changes by setting thresholds to obtain cell behavior trend analysis results;

[0122] First, the displacement data of all cells are analyzed to screen those cells with significant behavioral changes. The threshold method is used to determine whether the cell behavior shows characteristics of aggregation, division or migration. The significance of cell behavior is determined by setting a threshold. If the displacement speed of the cell in a specific time period exceeds the set threshold, it can be determined that the cell has migration behavior. If the displacement amplitude is small, it means that the cell is in a stationary or division stage. Based on this, cells can be divided into different groups and classified according to movement trends, and then the law of cell behavior changes is analyzed. For actual application scenarios, a specific displacement threshold (such as 10 pixels) can be set to distinguish between migration behavior and stationary behavior. If the cell displacement is greater than 10 pixels, it is considered that the cell exhibits migration behavior, otherwise it is division or stationary behavior, and finally the cell behavior trend analysis results are formed.

[0123] See also Figure 7 , the steps for obtaining the distance between cells and the corresponding speed are as follows:

[0124] Based on the results of cell behavior trend analysis, the positional relationship between cells is analyzed, the coordinate information of each pair of cells is extracted, the straight-line distance between cells is identified, the cell coordinates are extracted and the position changes of each pair of cells are tracked to obtain the distance data set between cells;

[0125] First, the position coordinates of the cells are extracted to obtain the coordinate points of the cells in each frame of the image. The coordinate points are automatically extracted through image processing technology, and the grayscale changes of the image pixels are used to calibrate the precise position of the cells. The straight-line distance between each pair of cells is calculated using the cell coordinates in adjacent frames. The distance calculation method uses the Euclidean distance formula to obtain the distance between each pair of cells. The coordinate difference between consecutive time frames can be used to track the corresponding position changes between cells, providing basic data for subsequent behavioral trend analysis. Assume that in the first frame, the coordinates of cell A are (100, 150) and the coordinates of cell B are (120, 180). In the second frame, the coordinates of cell A are (110, 160) and the coordinates of cell B are (125, 185). The distance change between cell A and cell B in the first and second frames is calculated, and the corresponding relative distance data between cells is calculated by the formula, and finally the distance data set between cells is obtained.

[0126] Based on the intercellular distance dataset, the intercellular velocity is calculated according to the displacement of each pair of cells in adjacent frames. The cell coordinate difference is derived to analyze the intercellular velocity change using the formula:

[0127]

[0128] Get the distance between cells and the corresponding speed;

[0129] Among them, v represents the velocity value of each pair of cells, D t+1 and D t is the intercellular distance between the t+1th frame and the tth frame, and Δt is the time interval;

[0130] The benefit of the formula is that by calculating the corresponding speeds between cells, the movement rate between cells can be quantified, and the dynamic behavior pattern of cells can be further analyzed;

[0131] v represents the corresponding speed between cells, D t+1 and D t are the inter-cell distances in the t+1th frame and the tth frame, respectively. Δt is the time interval. When calculating, the change in distance is divided by the time interval to obtain the corresponding velocity between cells, which helps to evaluate whether the cells are gathering or dispersing. Assume that in the 1st and 2nd frames, the distances between cell A and cell B are 10 and 15 pixels, respectively, and the time interval is 1 second.

[0132] Then the corresponding speed between cells is:

[0133] This result shows that the correspondence speed between cell A and cell B is 5 pixels / second, and the cells are dispersing rapidly. Further behavioral pattern analysis will infer whether the cells continue to disperse or tend to aggregate based on this speed, and obtain the distance between cells and the corresponding speed.

[0134] See also Figure 8 , the specific steps for obtaining the results of monitoring the coordinated behavior of cell populations are as follows:

[0135] Based on the distance and corresponding speed between cells, the spatial position coordinates and movement state of each cell are captured, the corresponding speed of the cells is monitored, the distance and speed of each cell with surrounding cells are recorded, and a cell group position and speed data set is generated;

[0136] First, the spatial position coordinates of each cell in the microscope image are determined through image recognition technology. The coordinates are converted from the pixel position of the cell in the image to the actual spatial position. The movement state of each cell is calculated by comparing consecutive image frames, and the cell's displacement and the change of displacement over time are calculated, and then its corresponding speed is obtained. In order to accurately monitor the corresponding speed between cells, the distance between each cell and the surrounding neighboring cells is calculated in real time. By comparing their respective spatial position coordinates, the straight-line distance between each pair of cells is obtained. The speed of each cell and the surrounding cells is calculated by tracking the change of their position over time, so as to obtain the cell's movement direction and speed. This involves the precise matching and trajectory tracking of cells between consecutive image frames, generating a cell population position and speed data set. The data set provides the basis for subsequent cell behavior analysis.

[0137] Based on the cell population position and velocity dataset, the corresponding position and motion state between each cell and its surrounding cells are analyzed, and cells with interaction strength greater than a set threshold are screened. The interaction range of each pair of cells is delineated based on distance, velocity, and interaction time, and a dataset of cell-cell interaction relationships is constructed.

[0138] First, the spatial position and velocity data of the cell population are used to calculate the corresponding position and motion state of each cell and its surrounding cells. The core of the analysis is to evaluate the relative motion direction and speed changes between cells. By calculating the corresponding speed of each pair of cells, it is evaluated whether they are approaching or moving away from each other. According to the preset interaction strength threshold, cell pairs with significant interactions are screened out. When the corresponding speed between cells is less than a certain value, it is considered that there is a strong interaction between the cells. Based on the screened cell pairs, combined with the spatial distance, speed difference and the duration of the interaction, the interaction range of each pair of cells is further delineated. The range is determined by calculating the changing trend of the interaction time and distance between cells, and constructing a cell-to-cell interaction relationship dataset. The dataset can be used to conduct a more in-depth analysis of the behavioral patterns and interactions between cells.

[0139] Using the intercellular interaction relationship dataset, we analyze the coordinated behavior of cell groups. By comparing the interactions between cells with group trends, we can determine the dynamic changes in group behavior and obtain the monitoring results of cell group coordinated behavior.

[0140] First, we conduct an in-depth analysis of the cell-to-cell interaction dataset to evaluate the intensity and time span of cell-to-cell interactions, thereby revealing the collaborative behavior patterns between cell groups. Through the distance between cells and the corresponding speed data, we identify closely cooperating cell pairs in the group, and then analyze how cells adjust their group behavior through interactions. The behavioral changes of the cell group are manifested as the strengthening or weakening of the cell-to-cell interaction. Based on the interaction information in the dataset, we can track the changing trend of the collaborative effect within the group. The interaction between cells depends not only on the physical distance, but also on the environmental factors inside and outside the cell. Therefore, dynamic monitoring of changes in cell behavior is crucial to understanding the health status of the cell group. Through continuous monitoring and data updates, we can accurately judge the dynamic changes in cell group behavior and generate monitoring results of cell group collaborative behavior.

[0141] The above are merely preferred embodiments of the present invention and do not limit the present invention in any other form. Any technician familiar with the profession may use the technical content disclosed above to change or modify it into an equivalent embodiment with equivalent changes and apply it to other fields. However, any simple modification, equivalent change and modification made to the above embodiment based on the technical essence of the present invention without departing from the content of the technical solution of the present invention shall still fall within the scope of protection of the technical solution of the present invention.

Claims

1. A CAR-T cell culture monitoring system based on image recognition, characterized in that: The system comprises: The cell region segmentation module analyzes the gradient intensity of cell culture images and divides the images into high-feature regions and low-feature regions. It then detects cell edges in high-feature regions, extracts edge positions and cell morphological features, refines cell shape and texture, and dynamically adjusts the segmentation threshold to obtain cell boundary segmentation results. The steps for obtaining the edge position and cell morphology characteristics are specifically as follows: Based on the cell culture image, the image is converted into a grayscale image, the brightness value of each pixel and the grayscale difference of the surrounding pixels are analyzed, the gradient value of each pixel is calculated, the gradient value is used to divide the image into regions, and the regional features are judged according to the gradient value to obtain the high and low feature region division results; Based on the high- and low-feature region division results, regional enhancement processing is performed on the pixels in the high-feature region, the local contrast of each pixel in the high-feature region is analyzed, and weights are assigned according to the difference in surrounding pixel values ​​to obtain enhanced data for the corresponding region; Based on the enhanced data of the corresponding area, the edge position of the cell is located by contour, the gradient change of each pixel in the high-feature area is analyzed, the edge pixel point is determined, and the cell morphology is analyzed according to the edge information to obtain the edge position and cell morphological characteristics; The steps for obtaining the cell boundary segmentation result are specifically as follows: Extracting edge information from the cell image based on the edge position and cell morphology characteristics, identifying the position information of each cell boundary, and generating a cell boundary position feature dataset by analyzing the cell morphology characteristics, including curvature and edge smoothness; By combining the cell boundary position feature dataset with cell texture information, texture features in the image are extracted, the changing pattern of texture features in the differentiated edge area is analyzed, and the segmentation threshold is dynamically adjusted to obtain a segmentation threshold adjustment coefficient; According to the segmentation threshold adjustment coefficient, the optimal matching of cell morphology and texture features is used, and the formula is adopted: Generate cell boundary segmentation results; Among them, T final represents the cell boundary segmentation threshold, T base represents the initial segmentation threshold, ω i Represents the weight coefficient, X i represents the adjustment coefficient of the association between cell morphology and texture features, and n is the number of features; The cell morphology analysis module analyzes the morphology of cells in high-feature areas based on the cell boundary segmentation results, extracts local texture information of cells, analyzes cell morphology differences through local feature comparison, refines the geometric and structural features of cells, and obtains a cell morphology feature dataset; The steps for obtaining the local texture information of the cell are specifically as follows: Based on the cell boundary segmentation results, boundary refinement and correction are performed, irregular edges are removed by local smoothing, the boundaries are repaired in combination with cell contour features, breaks and discontinuities are avoided, and morphology optimization is performed according to pixel distribution to obtain the corrected cell boundaries; Based on the corrected cell boundaries, geometric feature extraction is performed, by identifying the area and perimeter of each cell, using the boundary points to calculate the aspect ratio and shape factor, analyzing the symmetry and circularity of the cells, and screening the cell shape based on the area and perimeter ratio to obtain cell size and shape characteristics; Based on the cell size and shape characteristics, a local area is selected for texture analysis, pixel value changes in the local area of ​​the cell are counted, the degree of texture variation and consistency are quantified, the texture distribution pattern is analyzed in combination with the boundary information, and the texture difference and consistency between regions are evaluated to obtain the local texture information of the cell; The steps for obtaining the cell morphology feature dataset are specifically as follows: According to the local texture information of the cells, the cell edges, texture lines and contrast between regions are extracted, and combined with regional division, the image grayscale changes, local contrast, and texture direction parameters are identified, and the geometric differences between local regions are analyzed to obtain cell texture features; Based on the cell texture features, the cell morphology differences and geometric features between differentiated regions are quantified and compared with the texture similarity metric between regions using the formula: Obtain cell morphology difference data; Among them, δ represents the local texture difference, T r,j is the texture feature of the right area, T l,j is the texture feature of the left area, N is the number of texture points; Based on the cell morphology difference data, the geometric and structural features of the cells are refined through clustering and segmentation methods, and regions with similar geometric features are grouped using distance measurement judgment to obtain a cell morphology feature dataset; The cell behavior dynamic analysis module tracks and analyzes cell behavior based on the cell morphological feature dataset, monitors cell division, aggregation, and migration, evaluates distance and behavior trends through adjacent frame images, records cell dynamic behavior, analyzes dynamic changes in cell culture, and obtains cell behavior trend analysis results; The cell population collaborative monitoring module analyzes the synergistic effects between cell populations based on the cell behavior trend analysis results, monitors the trends of cell aggregation and dispersion, calculates the distance and corresponding speed between cells based on the relationship between cells, evaluates the overall behavior pattern of the cell population, dynamically monitors the synergistic effects of the cell population, and obtains the cell population collaborative behavior monitoring results.

2. The CAR-T cell culture monitoring system based on image recognition according to claim 1, characterized in that: The steps for obtaining the cell behavior trend analysis results are specifically as follows: Analyze each frame of the cell morphology feature dataset to extract cell position and identify the distance between adjacent frames, use the distance to judge cell division, aggregation and migration behavior, track cell position through time series and combined with image pixel position to obtain cell behavior tracking data; Based on the cell behavior tracking data, the dynamic behavior of the cells was evaluated according to the motion trajectory and change trend between adjacent frames, and the displacement distance value of the cells between adjacent frames was calculated using the formula: Obtain cell movement trend data; Among them, d represents the displacement distance of the cell between adjacent frames, x t+1 and y t+1 is the coordinate of the cell in frame t+1, x t and y t is the coordinate of the cell in frame t; Based on the cell movement trend data, the cell division, aggregation and migration behaviors are analyzed, the cell behavior trends are evaluated, and whether there is a changing trend is analyzed. The criticality of the cell behavior change is judged by setting a threshold to obtain the cell behavior trend analysis results.

3. The CAR-T cell culture monitoring system based on image recognition according to claim 2, characterized in that: The steps for obtaining the distance between cells and the corresponding speed are specifically as follows: Based on the cell behavior trend analysis results, analyzing the positional relationship between cells, extracting the coordinate information of each pair of cells, identifying the straight-line distance between cells, extracting the cell coordinates and tracking the position changes of each pair of cells, and obtaining a distance data set between cells; Based on the intercellular distance dataset, the intercellular velocity was calculated according to the displacement of each pair of cells in adjacent frames. The intercellular velocity change was analyzed by taking the derivative of the cell coordinate difference, using the formula: Get the distance between cells and the corresponding speed; Among them, v represents the velocity value of each pair of cells, D t+1 and D t is the intercellular distance between the t+1th frame and the tth frame, and Δt is the time interval.

4. The CAR-T cell culture monitoring system based on image recognition according to claim 3, characterized in that: The steps for obtaining the monitoring results of the cell population coordinated behavior are specifically as follows: Based on the distance and corresponding speed between the cells, the spatial position coordinates and motion state of each cell are captured, the corresponding speed of the cells is monitored, the distance and speed of each cell to the surrounding cells are recorded, and a cell group position and speed data set is generated; Based on the cell population position and velocity dataset, the corresponding position and motion state of each cell and its surrounding cells are analyzed, cells with interaction strength greater than a set threshold are screened, and the interaction range of each pair of cells is defined based on distance, velocity, and interaction time, thereby constructing a dataset of cell-to-cell interaction relationships. The intercellular interaction relationship dataset is used to analyze the coordinated behavior of the cell population. The dynamic changes of the group behavior are judged through the interaction between cells and the group trend, and the monitoring results of the cell group coordinated behavior are obtained.

Images