Artificial intelligence-based sperm cell classification statistical method and related equipment
The AI-based method for analyzing sperm cell images addresses the high cost and inaccuracy of existing DNA fragmentation detection methods by optimizing and segmenting images to accurately classify DNA fragmentation levels, providing a cost-effective and efficient solution for sperm cell analysis.
Patent Information
- Application Number
- CN202411674416.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-11-21
- Publication Date
- 2025-03-14
AI Technical Summary
Current methods for detecting DNA fragmentation in sperm cells, such as those using flow cytometry, are costly and lack accuracy, especially for distinguishing between different types of cellular debris, leading to inefficient and imprecise cell classification.
A method employing artificial intelligence to analyze fluorescently stained sperm cell images, converting them to grayscale, optimizing, and segmenting to identify cell contours, and then statistically categorizing DNA fragmentation levels without requiring expensive flow cytometry equipment.
This approach enhances image quality, improves detection accuracy, and reduces costs by enabling precise classification of DNA fragmentation levels in sperm cells, offering a more efficient and cost-effective solution for sperm cell analysis.
Smart Images

Figure CN119625725A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of cell statistics technology, and in particular to a sperm cell classification and statistics method based on artificial intelligence and related equipment. Background Art
[0002] The detection of sperm deoxyribonucleic acid (DNA) fragmentation is an important test item for evaluating male reproductive health. Its main principle is to evaluate sperm quality by detecting the integrity of sperm DNA. Under normal circumstances, sperm DNA should be intact, but in some cases, such as excessive oxidative stress, inflammation and other factors, sperm DNA may be damaged, broken or fragmented. These damages will affect the fertility of sperm and may even cause infertility.
[0003] The technology for detecting sperm DNA fragmentation index (DFI) on the market is mainly based on flow cytometer platform detection, but due to its high purchase and maintenance costs, most terminals that want to carry out DFI projects are discouraged due to cost and other issues. Analyzing the existing market, a new solution that can meet clinical testing needs and is stable and mature is urgently needed. In addition, the flow cytometer detection method is not ideal for distinguishing impurities in special samples, resulting in low efficiency and low accuracy in technical identification of cells, so other methods are needed to solve this problem.
[0004] It can be seen that the current sperm cell classification and counting method has the problem of high cost and inaccuracy in sperm cell classification and counting. Summary of the invention
[0005] The present application provides a sperm cell classification and statistics method and related equipment based on artificial intelligence, which can solve the problem of high cost and inaccuracy in sperm cell classification and statistics.
[0006] In a first aspect, an embodiment of the present application provides a sperm cell classification and statistics method based on artificial intelligence, and the sperm cell classification and statistics method includes:
[0007] Obtain a fluorescent staining image of the target sperm cell group, and perform grayscale conversion on the fluorescent staining image to obtain a grayscale image; sperm cells with various degrees of DNA fragmentation in the fluorescent staining image correspond to various colors one by one, and the various degrees of DNA fragmentation include severe fragmentation, moderate fragmentation, mild fragmentation and normal, severe fragmentation corresponds to red, moderate fragmentation corresponds to orange, mild fragmentation corresponds to yellow, and normal corresponds to green;
[0008] The grayscale image is optimized and adjusted to obtain an optimized image, and the optimized image is detected using a target detection model to obtain multiple cell contours; the multiple cell contours correspond to the multiple sperm cells one by one;
[0009] Segment the optimized image according to all cell contours to obtain multiple sub-images, and denoise all sub-images to obtain multiple target sub-images;
[0010] The color statistics of all cell outlines in all target sub-images were performed to obtain the number of sperm cells with various degrees of DNA fragmentation.
[0011] Optionally, the grayscale image is optimized and adjusted to obtain an optimized image, including:
[0012] Perform brightness equalization on the grayscale image to obtain a brightness equalized image;
[0013] Binarize the brightness balanced image to obtain a binary image;
[0014] Filtering the binary image to obtain a filtered image, and restoring the filtered image to obtain a restored image;
[0015] The restored image is denoised using a sliding window to obtain an optimized image.
[0016] Optionally, performing brightness equalization on the grayscale image to obtain a brightness equalized image includes:
[0017] Divide the pixel value into multiple intervals according to the size of the pixel value, and count the number of pixels corresponding to each pixel value interval in the grayscale image;
[0018] Based on the number of pixels corresponding to all pixel value intervals, the histogram equalization algorithm is used to equalize the brightness of the grayscale image to obtain a brightness-balanced image.
[0019] Optionally, binarizing the brightness-balanced image to obtain a binarized image includes:
[0020] For each pixel in the brightness balanced image, perform the following steps:
[0021] Determine whether the pixel value of the pixel point is greater than the pixel threshold;
[0022] If so, the pixel value of the pixel point is assigned the minimum pixel value;
[0023] Otherwise, the pixel value of the pixel is assigned to the maximum pixel value.
[0024] Optionally, the optimized image is segmented according to all cell outlines to obtain multiple sub-images, including:
[0025] Determine a plurality of target cell contours from all cell contours, and use the image portion corresponding to each target cell contour as the target image portion;
[0026] For each target image part, perform the following steps:
[0027] Calculating the similarity between the target image portion and the adjacent image portion, if the similarity is greater than or equal to a preset similarity value, taking both the adjacent image portion and the target image portion as image portions in a potential sub-image;
[0028] The adjacent image part is taken as the target image part, and the step of calculating the similarity between the target image part and the adjacent image part is returned until the similarity between the target image part and each adjacent image part is less than the preset similarity value, and the potential sub-image is taken as a sub-image.
[0029] Optionally, all sub-images are denoised to obtain multiple target sub-images, including:
[0030] For each sub-image respectively, if the area of the sub-image is greater than the area threshold, the sub-image is taken as a target sub-image.
[0031] Optionally, color statistics are performed on all cell outlines in all target sub-images to obtain the number of sperm cells with various degrees of DNA fragmentation, including:
[0032] According to the color corresponding to each cell outline in the fluorescent staining image in all target sub-images, the number of cell outlines corresponding to each color is counted;
[0033] For each color, the number of cell outlines corresponding to the color is taken as the number of sperm cells with DNA fragmentation of the degree corresponding to the color.
[0034] In a second aspect, the present application embodiment provides a sperm cell classification and statistics device based on artificial intelligence, comprising:
[0035] A grayscale conversion module is used to obtain a fluorescent staining image of a target sperm cell group and perform grayscale conversion on the fluorescent staining image to obtain a grayscale image; sperm cells with various degrees of DNA fragmentation in the fluorescent staining image correspond to various colors one by one, and the various degrees of DNA fragmentation include severe fragmentation, moderate fragmentation, mild fragmentation and normal, severe fragmentation corresponds to red, moderate fragmentation corresponds to orange, mild fragmentation corresponds to yellow, and normal corresponds to green;
[0036] An optimization and adjustment module is used to optimize and adjust the grayscale image to obtain an optimized image, and detect the optimized image using a target detection model to obtain multiple cell contours; the multiple cell contours correspond to the multiple sperm cells one by one;
[0037] A segmentation module is used to segment the optimized image according to all cell contours to obtain multiple sub-images, and to denoise all sub-images to obtain multiple target sub-images;
[0038] The color statistics module is used to perform color statistics on all cell outlines in all target sub-images to obtain the number of sperm cells with various degrees of DNA fragmentation.
[0039] In a third aspect, an embodiment of the present application provides a terminal device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the above-mentioned artificial intelligence-based sperm cell classification and statistical method when executing the above-mentioned computer program.
[0040] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, which stores a computer program, and when the computer program is executed by a processor, it implements the above-mentioned artificial intelligence-based sperm cell classification and statistical method.
[0041] The above solution of the present application has the following beneficial effects:
[0042] In an embodiment of the present application, a fluorescent staining image of a target sperm cell group is obtained, and the fluorescent staining image is gray-scale converted to obtain a gray-scale image, and then the gray-scale image is optimized and adjusted to obtain an optimized image, and the optimized image is detected using a target detection model to obtain multiple cell contours, and then the optimized image is segmented according to all cell contours to obtain multiple sub-images, and all sub-images are denoised to obtain multiple target sub-images, and finally all cell contours in all target sub-images are color counted to obtain the number of sperm cells with various degrees of DNA fragmentation. Among them, optimizing and adjusting the fluorescent staining image can improve the image quality of the fluorescent staining image, detecting high-quality images can improve the accuracy of detection, segmenting and denoising the detection results can further improve the accuracy of contour detection, and performing classification statistics based on accurate detection results improves the accuracy of classification statistics. The number of sperm cells with different degrees of DNA fragmentation is counted using fluorescent staining images, without the need to use expensive equipment such as flow cytometer platforms, which effectively reduces the cost of classification statistics.
[0043] Other beneficial effects of the present application will be described in detail in the subsequent specific implementation section. BRIEF DESCRIPTION OF THE DRAWINGS
[0044] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.
[0045] Figure 1 A flowchart of a sperm cell classification and statistical method based on artificial intelligence provided in one embodiment of the present application;
[0046] Figure 2 A schematic diagram of brightness balance provided in an embodiment of the present application;
[0047] Figure 3 A schematic diagram of a cell outline provided in an embodiment of the present application;
[0048] Figure 4 A schematic diagram of the structure of a sperm cell classification and statistics device based on artificial intelligence provided in one embodiment of the present application;
[0049] Figure 5 A schematic diagram of the structure of a terminal device provided in one embodiment of the present application. DETAILED DESCRIPTION
[0050] In the following description, specific details such as specific system structures, technologies, etc. are provided for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present application. However, it should be clear to those skilled in the art that the present application may also be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to prevent unnecessary details from obstructing the description of the present application.
[0051] It should be understood that when used in the present specification and the appended claims, the term "comprising" indicates the presence of described features, integers, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or combinations thereof.
[0052] It should also be understood that the term “and / or” used in the specification and appended claims refers to any and all possible combinations of one or more of the associated listed items, and includes these combinations.
[0053] As used in the specification and appended claims of this application, the term "if" can be interpreted as "when" or "uponce" or "in response to determining" or "in response to detecting", depending on the context. Similarly, the phrase "if it is determined" or "if [described condition or event] is detected" can be interpreted as meaning "uponce it is determined" or "in response to determining" or "uponce [described condition or event] is detected" or "in response to detecting [described condition or event]", depending on the context.
[0054] In addition, in the description of the present application specification and the appended claims, the terms "first", "second", "third", etc. are only used to distinguish the descriptions and cannot be understood as indicating or implying relative importance.
[0055] References to "one embodiment" or "some embodiments" etc. described in the specification of this application mean that one or more embodiments of the present application include specific features, structures or characteristics described in conjunction with the embodiment. Therefore, the statements "in one embodiment", "in some embodiments", "in some other embodiments", "in some other embodiments", etc. that appear in different places in this specification do not necessarily refer to the same embodiment, but mean "one or more but not all embodiments", unless otherwise specifically emphasized in other ways. The terms "including", "comprising", "having" and their variations all mean "including but not limited to", unless otherwise specifically emphasized in other ways.
[0056] In view of the high cost and inaccuracy of existing cell statistics, the embodiment of the present application provides a sperm cell classification and statistics method based on artificial intelligence, which obtains a fluorescent staining image of a target sperm cell group, performs grayscale conversion on the fluorescent staining image, obtains a grayscale image, then optimizes and adjusts the grayscale image to obtain an optimized image, and uses a target detection model to detect the optimized image to obtain multiple cell contours, and then segments the optimized image according to all cell contours to obtain multiple sub-images, and denoises all sub-images to obtain multiple target sub-images, and finally performs color statistics on all cell contours in all target sub-images to obtain the number of sperm cells with various degrees of DNA fragmentation. Among them, optimizing and adjusting the fluorescent staining image can improve the image quality of the fluorescent staining image, detecting high-quality images can improve the accuracy of detection, segmenting and denoising the detection results can further improve the accuracy of detection, and performing classification statistics based on accurate detection results improves the accuracy of classification statistics. The number of sperm cells with different degrees of DNA fragmentation is counted using fluorescent staining images, without the need to use expensive equipment such as flow cytometer platforms, which effectively reduces the cost of classification statistics.
[0057] Next, an exemplary description is given of the sperm cell classification and statistical method based on artificial intelligence provided in this application.
[0058] like Figure 1 As shown, the sperm cell classification and statistical method based on artificial intelligence provided in this application includes the following steps:
[0059] Step 11, obtaining a fluorescent staining image of the target sperm cell population, and performing grayscale conversion on the fluorescent staining image to obtain a grayscale image.
[0060] In the related art, a fluorescent color reagent is dripped into a target sperm cell group, and sperm cells with different degrees of DNA fragmentation exhibit different colors after actual reaction with the fluorescent color, which can present different color signals, and the morphology of the sperm cells can be clearly seen.
[0061] In the fluorescent staining image, sperm cells with various degrees of DNA fragmentation correspond to various colors one by one. Various degrees of DNA fragmentation include severe fragmentation, moderate fragmentation, mild fragmentation and normal. Severe fragmentation corresponds to red, moderate fragmentation corresponds to orange, mild fragmentation corresponds to yellow, and normal corresponds to green. The target sperm cell group includes multiple sperm cells that need to be classified and counted. The fluorescent staining image is a color image obtained by using a microscope or other equipment after fluorescent staining the target sperm cell group.
[0062] In some embodiments of the present application, computer software such as python can be used to perform grayscale conversion on the fluorescent staining image to obtain a grayscale image.
[0063] Step 12, optimizing and adjusting the grayscale image to obtain an optimized image, and detecting the optimized image using a target detection model to obtain multiple cell contours.
[0064] The multiple cell outlines correspond one-to-one to the multiple sperm cells.
[0065] In some embodiments of the present application, the above steps of optimizing and adjusting the grayscale image to obtain an optimized image, and detecting the optimized image using a target detection model to obtain multiple cell contours are specifically as follows:
[0066] The first step is to perform brightness equalization on the grayscale image to obtain a brightness equalized image.
[0067] First, the pixel value is divided into multiple intervals according to the size of the pixel value, and the number of pixels corresponding to each pixel value interval in the grayscale image is counted.
[0068] Then, based on the number of pixels corresponding to all pixel value intervals, the histogram equalization algorithm is used to perform brightness equalization on the grayscale image to obtain a brightness balanced image.
[0069] For example, the histogram of a grayscale image is Figure 2 As shown in the figure, the horizontal axis represents the pixel value and the vertical axis represents the number of pixels. Figure 2 a is the histogram of the grayscale image before brightness equalization. The maximum value of the grayscale image pixel is 256. The pixel value histogram of the grayscale image is divided into 256 levels. At the same time, 8 image grayscale segmentation intervals (i.e. pixel value intervals) are defined, and the segmentation values are 0, 40, 80, 120, 160, 200, 220, and 255. Traverse the pixels in each row and column of the grayscale image, put the pixels with pixel values between 0 and 40 in one interval, 40 and 80 in another interval, and so on, and record the number of pixels in each interval. Observation Figure 2 a. If the difference between the maximum and minimum values in the vertical axis direction exceeds 1000, it means that the image brightness is uneven, and brightness equalization processing is required. The pixel values counted in the 8 intervals are normalized, and the proportion of pixels in the interval to the total number of pixels in the entire image is calculated. The proportions within the 8 intervals are accumulated respectively. If the value obtained during the accumulation process is close to the value of 0-7 (8 intervals), it is recorded as a level. The first interval 0 represents 790 pixels between 0 and 40, accounting for 0.19 of the total pixels. 0.19 and 1 / 7 are similar, so they are recorded as level s0. Add 0.19 of interval 0 and 0.25 of interval 1 to get 0.44, which is similar to 3 / 7 and is recorded as s1. After adding them one by one, when the accumulated result is the same as rk, they are all recorded as the same level. Finally, the specific number of pixels in the ratio s0-s4 is counted, and the pixel value of each pixel is recalculated according to the histogram equalization algorithm. The pixel value is redistributed to the original grayscale image to obtain the processed histogram (such as Figure 2 b), achieving the effect of brightness equalization.
[0070] The second step is to binarize the brightness balanced image to obtain a binary image.
[0071] Specifically, for each pixel in the brightness balanced image, perform the following steps:
[0072] Determine whether the pixel value of the pixel is greater than the pixel threshold.
[0073] If so, the pixel value of the pixel is assigned to the minimum pixel value.
[0074] Otherwise, the pixel value of the pixel is assigned to the maximum pixel value.
[0075] Exemplarily, the pixel threshold is 127, the maximum pixel value is 255, the minimum pixel value is 0, and if the pixel value is greater than 127, the pixel value in the image is set to 0, and if the pixel value is less than 127, the pixel value in the image is set to 255.
[0076] The third step is to filter the binary image to obtain a filtered image, and restore the filtered image to obtain a restored image.
[0077] In some embodiments of the present application, an adaptive non-local mean filter is used to dynamically adjust the filter parameters according to the local texture and edge features of the binary image, remove noise and retain the sperm structure information in the binary image. Then, a multi-scale wavelet transform is applied to perform targeted filtering on noise components of different frequencies, while retaining the high-frequency details of the binary image to obtain a filtered image. Then, an energy-based image restoration technique is used, combined with the neighborhood information of the filtered image, to optimize the denoising effect, and sparse representation and beamforming techniques are introduced to extract the sparse representation of the filtered image, and remove noise by simulating beam focusing, restore the details of the filtered image, and obtain a restored image. In addition, the filter parameters of the adaptive non-local mean filter can be adjusted according to the local statistical characteristics of the restored image in combination with the analysis of the adaptive filtering and image content, so that the filter can better adapt to the characteristics of different regions, thereby improving the noise removal effect.
[0078] The fourth step is to use the sliding window to denoise the restored image to obtain the optimized image.
[0079] Exemplarily, a 5*5 sliding window is used, and the pixels in the sliding window are scanned from top to bottom and from left to right, and the obtained pixel values are arranged in ascending order according to the bubble sort algorithm, and the middle pixel is taken, and the obtained middle pixel is filled into the pixel corresponding to the center position of the sliding window, and the scanning and replacement are performed in sequence to finally obtain an optimized image.
[0080] In the fifth step, the optimized image is detected using the target detection model to obtain multiple cell contours.
[0081] Specifically, the optimized image is detected using the target detection model to obtain multiple cell detection frames, where the cell detection frame is the circumscribed rectangle of the cell, and then each cell detection frame is processed to obtain the cell outline.
[0082] Exemplarily, the above-mentioned target detection model includes a detection model and a position reasoning model connected in sequence. The detection model can be a yolov8 model, which is used to detect the cell detection frame in the image. The position reasoning model can be an open neural network exchange model (ONNX, OpenNeural Network Exchange), which is used to infer the position of the detected cell detection frame. The cell detection frame is detected using the yolov8 model, and then the position of the cell detection frame is obtained using the ONNX model. Finally, the cell detection frame can be processed using algorithms such as opencv to obtain the cell outline.
[0083] It is worth mentioning that optimizing and adjusting the fluorescent staining images can improve the image quality of the fluorescent staining images, and detecting high-quality images can improve the accuracy of detection.
[0084] Step 13, segmenting the optimized image according to all cell contours to obtain multiple sub-images, and denoising all sub-images to obtain multiple target sub-images.
[0085] In some embodiments of the present application, the above steps of segmenting the optimized image according to all cell contours to obtain multiple sub-images, and denoising all sub-images to obtain multiple target sub-images are specifically as follows:
[0086] In the first step, multiple target cell contours are determined from all cell contours, and the image portion corresponding to each target cell contour is used as the target image portion.
[0087] In some embodiments of the present application, multiple target cell contours may be determined manually. The target cell contours are manually selected cell contours with clear contours.
[0088] In the second step, for each target image part, perform the following steps:
[0089] First, the similarity between the target image part and the adjacent image part is calculated, and if the similarity is greater than or equal to a preset similarity value, the adjacent image part and the target image part are both regarded as image parts in a potential sub-image. That is, the adjacent image part and the target image part are combined into a potential sub-image.
[0090] Then, the adjacent image part is taken as the target image part, and the step of calculating the similarity between the target image part and the adjacent image part is returned until the similarity between the target image part and each adjacent image part is less than the preset similarity value, and the potential sub-image is taken as a sub-image.
[0091] It should be noted that the sub-image obtained in the above manner includes at least one cell outline.
[0092] Exemplarily, the above similarity can be calculated using the cosine similarity calculation formula. The optimized image can also be segmented using an image segmentation algorithm based on region growing to obtain multiple sub-images. If the similarity between the target image part a and the adjacent image part b is greater than or equal to the similarity threshold, the target image part a and the adjacent image part b are both used as image parts in a potential sub-image, and the adjacent image part b is used as the target image part b. At this time, the similarity between the target image part b and the adjacent image part c is greater than or equal to the similarity threshold, then the adjacent image part c is also merged into the potential sub-image. At this time, the potential sub-image includes three image parts a, b, and c.
[0093] In the third step, for each sub-image, if the area of the sub-image is greater than the area threshold, the sub-image is regarded as a target sub-image.
[0094] In some embodiments of the present application, if the area of the sub-image is less than or equal to the area threshold, the sub-image is removed. It is also possible to determine whether the cell contour is noise by analyzing the roundness, convexity and other parameters of the cell contour in the sub-image. For example, when the roundness of the cell contour is less than the roundness threshold and the convexity is less than the convexity threshold, the cell contour is determined to be noise. When the roundness of the cell contour is greater than or equal to the roundness threshold or the convexity is greater than or equal to the convexity threshold, the cell contour is retained and the cell contour determined to be noise is removed to achieve the purpose of denoising.
[0095] It is worth mentioning that segmenting and denoising the detection results can further improve the accuracy of detection.
[0096] Step 14, performing color statistics on all cell outlines in all target sub-images to obtain the number of sperm cells with various degrees of DNA fragmentation.
[0097] In some embodiments of the present application, the above step of performing color statistics on all cell outlines in all target sub-images to obtain the number of sperm cells with various degrees of DNA fragmentation is specifically as follows:
[0098] In the first step, according to the color corresponding to each cell outline in the fluorescent staining image in all target sub-images, the number of cell outlines corresponding to each color is counted.
[0099] Specifically, the color of each pixel in the image area corresponding to each cell outline in the fluorescent staining image is obtained, and the color with the maximum proportion of all pixels is calculated, and the color is used as the color of the cell outline. For example, in the image area corresponding to the cell outline, red pixels account for 20%, orange pixels account for 76%, yellow pixels account for 2%, and green pixels account for 2%, then the color of the cell outline is orange. Then, according to the color of each cell outline, the number of cell outlines corresponding to each color is counted.
[0100] For example, the cell outline is as follows Figure 3 As shown in the figure, the ellipse represents the detected cell outline, the box represents the external matrix of the cell outline, and the different fillings of the ellipse represent different colors of the cell outline in the fluorescent staining image, corresponding to cells with different degrees of DNA fragmentation.
[0101] In the second step, for each color, the number of cell outlines corresponding to the color is taken as the number of sperm cells with DNA fragmentation of the degree corresponding to the color.
[0102] For example, after obtaining the number of sperm cells with each degree of DNA fragmentation, the proportion of cells with each degree of DNA fragmentation can be calculated, and the proportion reflects to a certain extent the degree of damage to sperm cells in the target sperm cell population.
[0103] It is worth mentioning that optimizing and adjusting the fluorescent staining images can improve the image quality of the fluorescent staining images, detecting high-quality images can improve the accuracy of detection, segmenting and denoising the detection results can further improve the accuracy of detection, and classifying and counting based on accurate detection results improves the accuracy of classification statistics. The number of sperm cells with different degrees of DNA fragmentation is counted using fluorescent staining images without the need to use expensive equipment such as flow cytometer platforms, which effectively reduces the cost of classification statistics. And the statistical method of the present application effectively improves the statistical efficiency compared to traditional manual detection and counting.
[0104] The following is an exemplary description of the sperm cell classification and statistics device based on artificial intelligence provided in this application.
[0105] like Figure 4 As shown, the embodiment of the present application provides a sperm cell classification and statistics device based on artificial intelligence, and the sperm cell classification and statistics device based on artificial intelligence 400 includes:
[0106] Grayscale conversion module 401 is used to obtain a fluorescent staining image of a target sperm cell group, and grayscale convert the fluorescent staining image to obtain a grayscale image; sperm cells with various degrees of DNA fragmentation in the fluorescent staining image correspond to a plurality of colors one by one, and the various degrees of DNA fragmentation include severe fragmentation, moderate fragmentation, mild fragmentation and normal, severe fragmentation corresponds to red, moderate fragmentation corresponds to orange, mild fragmentation corresponds to yellow, and normal corresponds to green;
[0107] The optimization and adjustment module 402 is used to optimize and adjust the grayscale image to obtain an optimized image, and detect the optimized image using a target detection model to obtain a plurality of cell contours; the plurality of cell contours correspond to the plurality of sperm cells one by one;
[0108] The segmentation module 403 is used to segment the optimized image according to all cell contours to obtain multiple sub-images, and to denoise all sub-images to obtain multiple target sub-images;
[0109] The color statistics module 404 is used to perform color statistics on all cell outlines in all target sub-images to obtain the number of sperm cells with various degrees of DNA fragmentation.
[0110] It should be noted that the information interaction, execution process, etc. between the above-mentioned devices / units are based on the same concept as the method embodiment of the present application. Their specific functions and technical effects can be found in the method embodiment part and will not be repeated here.
[0111] The technicians in the relevant field can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example for illustration. In practical applications, the above-mentioned function allocation can be completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiment can be integrated in a processing unit, or each unit can exist physically separately, or two or more units can be integrated in one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of software functional units. In addition, the specific names of the functional units and modules are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of this application. The specific working process of the units and modules in the above-mentioned system can refer to the corresponding process in the aforementioned method embodiment, which will not be repeated here.
[0112] like Figure 5 As shown, an embodiment of the present application provides a terminal device. The terminal device D10 of this embodiment includes: at least one processor D100 ( Figure 5 Only one processor is shown in the figure), a memory D101, and a computer program D102 stored in the memory D101 and executable on the at least one processor D100, wherein the processor D100 implements the steps of any of the above-mentioned method embodiments when executing the computer program D102.
[0113] Specifically, when the processor D100 executes the computer program D102, it obtains the fluorescent staining image of the target sperm cell group, performs grayscale conversion on the fluorescent staining image, obtains a grayscale image, then optimizes and adjusts the grayscale image to obtain an optimized image, and uses the target detection model to detect the optimized image to obtain multiple cell contours, and then segments the optimized image according to all cell contours to obtain multiple sub-images, and denoises all sub-images to obtain multiple target sub-images, and finally performs color statistics on all cell contours in all target sub-images to obtain the number of sperm cells with various degrees of DNA fragmentation. Among them, optimizing and adjusting the fluorescent staining image can improve the image quality of the fluorescent staining image, detecting high-quality images can improve the accuracy of detection, segmenting and denoising the detection results can further improve the accuracy of detection, and performing classification statistics based on accurate detection results improves the accuracy of classification statistics. The number of sperm cells with different degrees of DNA fragmentation is counted using fluorescent staining images, without the need to use expensive equipment such as flow cytometer platforms, which effectively reduces the cost of classification statistics.
[0114] The processor D100 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc.
[0115] In some embodiments, the memory D101 may be an internal storage unit of the terminal device D10, such as a hard disk or memory of the terminal device D10. In other embodiments, the memory D101 may also be an external storage device of the terminal device D10, such as a plug-in hard disk, a smart memory card (SMC, SmartMedia Card), a secure digital (SD, Secure Digital) card, a flash card (Flash Card), etc. equipped on the terminal device D10. Further, the memory D101 may also include both an internal storage unit of the terminal device D10 and an external storage device. The memory D101 is used to store an operating system, an application program, a boot loader (BootLoader), data and other programs, such as the program code of the computer program, etc. The memory D101 may also be used to temporarily store data that has been output or is to be output.
[0116] An embodiment of the present application further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps in the above-mentioned method embodiments can be implemented.
[0117] An embodiment of the present application provides a computer program product. When the computer program product runs on a terminal device, the terminal device can implement the steps in the above-mentioned method embodiments when executing the computer program product.
[0118] If the 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 storage medium. Based on this understanding, the present application implements all or part of the process in the above-mentioned embodiment method, which can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a computer-readable storage medium, and the computer program can implement the steps of the above-mentioned various method embodiments when executed by the processor. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may at least include: any entity or device, recording medium, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, RandomAccess Memory), electric carrier signal, telecommunication signal and software distribution medium that can carry the computer program code to the sperm cell classification and statistical method device / terminal device based on artificial intelligence. For example, a USB flash drive, a mobile hard disk, a magnetic disk or an optical disk. In some jurisdictions, according to legislation and patent practice, computer-readable media cannot be electric carrier signals and telecommunication signals.
[0119] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described or recorded in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.
[0120] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.
[0121] The above is a preferred embodiment of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles described in the present application. These improvements and modifications should also be regarded as the scope of protection of the present application.
Claims
1. A sperm cell classification and statistical method based on artificial intelligence, characterized in that: include: Acquiring a fluorescent staining image of a target sperm cell population, and performing grayscale conversion on the fluorescent staining image to obtain a grayscale image; The sperm cells with various degrees of DNA fragmentation in the fluorescent staining image correspond to various colors one by one, and the various degrees of DNA fragmentation include severe fragmentation, moderate fragmentation, mild fragmentation and normal, the severe fragmentation corresponds to red, the moderate fragmentation corresponds to orange, the mild fragmentation corresponds to yellow, and the normal corresponds to green; Optimizing and adjusting the grayscale image to obtain an optimized image, and detecting the optimized image using a target detection model to obtain a plurality of cell contours; the plurality of cell contours correspond one to one to the plurality of sperm cells; Segmenting the optimized image according to all cell contours to obtain a plurality of sub-images, and denoising all the sub-images to obtain a plurality of target sub-images; The color statistics of all cell outlines in all target sub-images were performed to obtain the number of sperm cells with various degrees of DNA fragmentation.
2. The sperm cell classification and statistical method according to claim 1, characterized in that: The step of optimizing and adjusting the grayscale image to obtain an optimized image includes: Performing brightness equalization on the grayscale image to obtain a brightness equalized image; Binarizing the brightness-balanced image to obtain a binary image; Filtering the binary image to obtain a filtered image, and restoring the filtered image to obtain a restored image; The restored image is denoised using a sliding window to obtain an optimized image.
3. The sperm cell classification and statistical method according to claim 2, characterized in that: The step of performing brightness equalization on the grayscale image to obtain a brightness equalized image includes: Dividing the image into multiple pixel value intervals according to the size of the pixel value, and counting the number of pixel points corresponding to each pixel value interval in the grayscale image; Based on the number of pixels corresponding to all pixel value intervals, a histogram equalization algorithm is used to perform brightness equalization on the grayscale image to obtain a brightness equalized image.
4. The sperm cell classification and statistical method according to claim 2, characterized in that: The step of binarizing the brightness balanced image to obtain a binarized image includes: For each pixel in the brightness balanced image, perform the following steps: Determine whether the pixel value of the pixel point is greater than a pixel threshold; If yes, assign the pixel value of the pixel point to the minimum pixel value; Otherwise, the pixel value of the pixel point is assigned the maximum pixel value.
5. The sperm cell classification and statistical method according to claim 1, characterized in that: The optimized image is segmented according to all cell contours to obtain a plurality of sub-images, including: Determine a plurality of target cell contours from all cell contours, and use an image portion corresponding to each of the target cell contours as a target image portion; For each target image part, perform the following steps: Calculating the similarity between the target image portion and the adjacent image portion, and if the similarity is greater than or equal to a preset similarity value, treating both the adjacent image portion and the target image portion as image portions in a potential sub-image; The adjacent image part is taken as the target image part, and the step of calculating the similarity between the target image part and the adjacent image part is returned until the similarity between the target image part and each adjacent image part is less than a preset similarity value, and the potential sub-image is taken as a sub-image.
6. The sperm cell classification and statistical method according to claim 1, characterized in that: The denoising of all sub-images is performed to obtain multiple target sub-images, including: For each of the sub-images, if the area of the sub-image is greater than an area threshold, the sub-image is taken as a target sub-image.
7. The sperm cell classification and statistical method according to claim 1, characterized in that: The color statistics of all cell outlines in all target sub-images are performed to obtain the number of sperm cells with various degrees of DNA fragmentation, including: According to the color corresponding to each cell outline in the fluorescent staining image in all target sub-images, the number of cell outlines corresponding to each color is counted; For each color, the number of cell outlines corresponding to the color is taken as the number of sperm cells with DNA fragmentation of the degree corresponding to the color.
8. A sperm cell classification and counting device based on artificial intelligence, comprising: A grayscale conversion module is used to obtain a fluorescent staining image of a target sperm cell group and perform grayscale conversion on the fluorescent staining image to obtain a grayscale image; The sperm cells with various degrees of DNA fragmentation in the fluorescent staining image correspond to various colors one by one, and the various degrees of DNA fragmentation include severe fragmentation, moderate fragmentation, mild fragmentation and normal, the severe fragmentation corresponds to red, the moderate fragmentation corresponds to orange, the mild fragmentation corresponds to yellow, and the normal corresponds to green; An optimization and adjustment module is used to optimize and adjust the grayscale image to obtain an optimized image, and detect the optimized image using a target detection model to obtain a plurality of cell contours; the plurality of cell contours correspond to the plurality of sperm cells one by one; A segmentation module, used for segmenting the optimized image according to all cell contours to obtain multiple sub-images, and denoising all sub-images to obtain multiple target sub-images; The color statistics module is used to perform color statistics on all cell outlines in all target sub-images to obtain the number of sperm cells with various degrees of DNA fragmentation.
9. A terminal device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the sperm cell classification and statistical method based on artificial intelligence as described in any one of claims 1 to 7 is implemented.
10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the artificial intelligence-based sperm cell classification and statistical method as described in any one of claims 1 to 7 is implemented.
Citation Information
Cited By
Processing method based on sperm cell detection and storage medium
CN121121402A
