A method and system for detecting, identifying and processing defective morphology of induced pluripotent stem cells in vitro

Abstract

The present invention discloses a method and system for detecting, identifying and processing defect morphology of induced pluripotent stem cells in vitro, wherein the method comprises: obtaining an image of a sperm head to be detected and an image of a sample sperm head; acquiring the contour of the sperm head image to be detected to obtain the contour of the head to be detected; acquiring the contour of the sample sperm head image to obtain the contour of the head sample; defect matching the contour of the head to be detected and the contour of the head sample, outputting the defect matching result, and determining the defect information according to the defect matching result.

Description

Technical Field

[0001] The present invention relates to the field of cell detection, and in particular to a method and system for detecting, identifying and processing defective morphology of induced pluripotent stem cells in vitro. Background Art

[0002] Sperm cells are cultured using stem cell technology, which mainly includes two treatment methods: in vitro induced pluripotent stem cells (iPSCs) and embryonic stem cells (ESCs).

[0003] In vitro induced pluripotent stem cells are cells obtained from the somatic cells of adults (currently men), which are reprogrammed to return to a state similar to embryonic stem cells, and then induced to differentiate into sperm cells through specific culture conditions.

[0004] Although stem cell technology has made some progress in culturing sperm cells, there are still some challenges and limitations. For example, the cultured sperm cells may have quality problems. Studies have found that there is a high rate of male infertility in the world. Male sperm is composed of three parts, namely the sperm head, the middle part of the sperm (i.e. the body), and the sperm tail (i.e. the flagellum). Studies have also found that if the quality of the man's sperm cells is poor, the quality of the sperm cells obtained by in vitro induced pluripotent stem cells is not high either. Therefore, it is necessary to determine the survival rate of sperm obtained by in vitro induced pluripotent stem cells by testing the characteristics of different parts, so as to verify the quality of sperm cells obtained by in vitro induced pluripotent stem cell culture.

[0005] The study also found that the sperm head is the most susceptible to deformities among the three parts due to the influence of reproductive system infection, varicocele and bad living habits (such as smoking and alcoholism); such as giant sperm (i.e., large-headed sperm), micro sperm (i.e., small-headed sperm), double-headed sperm, and long-headed sperm (i.e., pear-shaped sperm); due to the different deformities of the sperm head, there will be great difficulties in the detection process, and it is impossible to determine the deformity of the sperm head obtained by in vitro induced pluripotent stem cell culture, and the detection accuracy is low. In particular, the accuracy of the detection of different forms of sperm head deformities is low. Summary of the invention

[0006] The purpose of the present invention is to provide a method and system for detecting and identifying defective morphology of induced pluripotent stem cells in vitro, which solves the above-mentioned technical problems pointed out in the prior art. Therefore, the present invention detects different shapes of sperm heads through a method for detecting and identifying defective morphology of induced pluripotent stem cells in vitro.

[0007] The present invention provides a method for detecting, identifying and processing defective morphology of induced pluripotent stem cells in vitro, comprising the following specific operating steps:

[0008] Acquire the image of the sperm head to be detected and the image of the sample sperm head;

[0009] Acquiring the contour of the sperm head image to be detected to obtain the contour of the head to be detected;

[0010] Acquiring the contour of the sample sperm head image to obtain a head sample contour;

[0011] Defect matching is performed on the head contour to be detected and the head sample contour, a defect matching result is output, and defect information is determined according to the defect matching result.

[0012] Accordingly, the contour of the sperm head image to be detected is acquired to obtain the contour of the head to be detected. The specific operation steps are as follows:

[0013] Preprocessing the sperm head image to be detected to obtain a sperm grayscale image to be detected;

[0014] Calculating grayscale values ​​of pixels of the grayscale image of sperm to be detected; clustering significant pixels in the grayscale image of sperm to be detected to obtain a plurality of head pixel clusters to be detected;

[0015] Execute the judgment of whether the current head pixel cluster to be detected and the adjacent head pixel cluster to be detected belong to the same sperm cell: calculate the Euclidean distance between the head pixel cluster to be detected and the adjacent head pixel cluster to be detected, and traverse the Euclidean distance between all two adjacent head pixel clusters to be detected; preset the Euclidean distance threshold w, and judge whether the Euclidean distance between all two adjacent head pixel clusters to be detected is less than the preset Euclidean distance threshold w;

[0016] If not, the head pixel cluster to be detected in the neighborhood is eliminated; if so, it is determined that the current head pixel cluster to be detected and the adjacent head pixel cluster to be detected belong to the same sperm cell, and the current head pixel cluster to be detected and the adjacent head pixel cluster to be detected are obtained as the head target contour points to be detected, and all the confirmed head target contour points to be detected are connected to obtain the head contour to be detected.

[0017] Accordingly, the contour of the sample sperm head image is acquired to obtain the head sample contour. The specific operation steps are as follows:

[0018] Preprocessing the sample sperm head image to obtain a sample sperm grayscale image;

[0019] Calculating grayscale values ​​of pixel points of the sample sperm grayscale image; clustering significant pixel points in the sample sperm grayscale image to obtain a plurality of head sample pixel clusters;

[0020] Traversing the pixel points inside the head sample pixel cluster to calculate the average grayscale value, and obtaining the average grayscale value of the head sample pixel cluster; calculating the grayscale mean of the average grayscale values ​​of all the head sample pixel clusters, assuming that the grayscale mean of the average grayscale values ​​of all the head sample pixel clusters is a threshold value t, and judging whether the average grayscale value of the head sample pixel cluster is greater than the grayscale mean of the average grayscale values ​​of all the head sample pixel clusters as the threshold value t;

[0021] If not, the head sample pixel cluster is eliminated;

[0022] If so, the head sample pixel cluster is retained and confirmed as the head sample target contour point, and all the retained head sample target contour points are connected to obtain the head sample contour.

[0023] Accordingly, defect matching is performed on the head contour to be detected and the head sample contour, a defect matching result is output, and defect information is determined according to the defect matching result. The specific operation steps are as follows:

[0024] The contour of the head to be detected and the contour of the head sample are enclosed by a bounding box to determine a circular bounding box with a minimum radius, and the circular area of ​​the head to be detected and the circular area of ​​the head sample are obtained; by comparing the circular area of ​​the head to be detected and the circular area of ​​the head sample, the defect matching result of the sperm head image to be detected is obtained; and the defect information is determined according to the defect matching result.

[0025] Accordingly, the head contour to be detected and the head sample contour are bounded by a bounding box to determine a circular bounding box with a minimum radius, and the circular area of ​​the head to be detected and the circular area of ​​the head sample are obtained;

[0026] Randomly select two edge pixel points on the head contour to be detected in a minimum enclosing circle manner, connect the two edge pixel points with a straight line, and obtain a point-to-point straight line;

[0027] The two edge pixel points are two opposite points on the head contour to be detected, and after selecting the first edge pixel point, the edge pixel point on the head contour to be detected that is farthest from the first edge pixel point is selected as the second edge pixel point, and the first edge pixel point and the second edge pixel point are used as two opposite points;

[0028] Calculate the vertical distance between the remaining edge pixel points on the head contour to be detected and the point straight line, preset a vertical distance threshold j, and determine whether the current vertical distance is greater than the preset vertical distance threshold j;

[0029] If so, retain the edge pixel points for calculating the vertical distance;

[0030] If not, the edge pixels for calculating the vertical distance are eliminated;

[0031] Traverse all the remaining edge pixel points on the head contour to be detected to see if the vertical distance to the point straight line is greater than the preset vertical distance threshold j;

[0032] Selecting two edge pixel points with the farthest distance from each other from the retained edge pixel points as diameters to form a minimum bounding box, and calculating the area of ​​the minimum bounding box as the circular area of ​​the head to be detected;

[0033] The above steps are performed on the head sample contour to obtain the head sample circular area.

[0034] Accordingly, by comparing the circular area of ​​the head to be detected with the circular area of ​​the head sample, the defect matching result of the sperm head image to be detected is obtained. The specific operation steps are as follows:

[0035] Match the circular area of ​​the head to be detected with the circular area of ​​the head sample, and calculate the circular area difference; preset a circular area threshold u, and determine whether the circular area difference is less than the preset circular area threshold u;

[0036] If not, it is determined that the head image to be detected of the current sperm head image to be detected has defects, and the specific defect information is confirmed through the defect matching result of the sperm head image to be detected;

[0037] If so, the head image to be detected of the sperm head image to be detected has no deformity or variation.

[0038] Accordingly, the specific defect information is confirmed by the defect matching result of the sperm head image to be detected, and the specific operation steps are as follows:

[0039] When it is determined that the head image to be detected of the sperm head image to be detected has defects, a first defect threshold g and a second defect threshold h are preset, and a magnitude relationship between the circular area difference value in the first defect threshold g and the second defect threshold h is determined;

[0040] The first defect threshold g<the second defect threshold h;

[0041] If the circular area difference is less than the first defect threshold g, the defect information of the sperm head image to be detected is micro sperm;

[0042] If the circular area difference = the first defect threshold g, then the defect information of the sperm head image to be detected is pear-shaped sperm;

[0043] If the circular area difference is greater than the second defect threshold h, the defect information of the sperm head image to be detected is a large-headed sperm;

[0044] If the circular area difference = the second defect threshold h, the defect information of the sperm head image to be detected is a double-headed sperm;

[0045] If the first defect threshold g<circular area difference<second defect threshold h, then the sperm head image to be detected has no defect problem.

[0046] Accordingly, the present invention also proposes an in vitro induced pluripotent stem cell defect morphology detection and recognition processing system, comprising: an image acquisition module, a contour acquisition module, and a defect detection module;

[0047] The image acquisition module is used to acquire the image of the sperm head to be detected and the image of the sample sperm head;

[0048] The contour acquisition module is used to acquire the contour of the sperm head image to be detected, so as to obtain the contour of the head to be detected;

[0049] Acquiring the contour of the sample sperm head image to obtain a head sample contour;

[0050] The defect detection module is used to perform defect matching on the head contour to be detected and the head sample contour, output a defect matching result, and determine defect information according to the defect matching result.

[0051] Compared with the prior art, the embodiments of the present invention have at least the following technical advantages:

[0052] From the analysis of the above-mentioned in vitro induced pluripotent stem cell defect morphology detection and identification processing method provided by the present invention, it can be known that in specific applications, the sperm head image to be detected is converted into a sperm grayscale image to be detected to obtain a sperm grayscale image to be detected, and the grayscale value of the pixel point is calculated. The grayscale value of the pixel point of the sperm head inside the sperm grayscale image to be detected and the pixel point of the background image of the sperm grayscale image to be detected will be different, and the grayscale value of the pixel point of the sperm head will be very high. Therefore, the pixel points with high grayscale values ​​are clustered to obtain a number of head pixel clusters to be detected; the Euclidean distance between the head pixel clusters to be detected is calculated to determine whether the head pixel clusters to be detected belong to the same sperm. The head cell may still belong to two sperm head cells; the qualified head pixel clusters to be detected are confirmed as the head target contour points to be detected and connected to obtain the head target contour points to be detected; the sample sperm head image is a complete image, so the pixel points with high grayscale values ​​of the sample sperm head image are clustered to obtain the head sample pixel cluster; the pixel points inside the head sample pixel cluster are screened again by threshold value to obtain the head sample pixel cluster with high average grayscale value, which is confirmed as the head sample target contour point, so that clear head sample target contour points are screened twice, and a clear head sample contour is connected by clear head sample target contour points;

[0053] Further, two random edge pixel points of the head contour to be detected are selected by the minimum enclosing circle method, and the two edge pixel points are connected to obtain a point-to-point straight line; the vertical distances between the remaining edge pixel points and the point-to-point straight line are determined, and the non-conforming edge pixel points are eliminated, and the retained edge pixel points are used to further screen out the two conforming edge pixel points to form a diameter, and the minimum enclosing box is obtained by the diameter, and the circular area of ​​the head to be detected is further obtained; the circular area of ​​the head sample contour is obtained through the above steps;

[0054] The circular area of ​​the head to be detected is further compared with the circular area of ​​the head sample, and the circular area difference is used to determine whether the image of the sperm head to be detected has defects; if the circular area difference is smaller than the preset circular area threshold, it means that the circular area of ​​the head to be detected and the circular area of ​​the head sample are close to each other, and if they are close to each other, the image of the sperm head to be detected has no defects; if it is larger than the preset circular area threshold, it means that the image of the sperm head to be detected has defects, and the defect information of the sperm head image to be detected is determined by the defects. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0056] Figure 1 This is an overall flow chart of a method for detecting and identifying defective morphology of in vitro induced pluripotent stem cells provided in Example 1 of the present invention;

[0057] Figure 2 A flowchart of obtaining a profile of a method for detecting and identifying defective morphology of in vitro induced pluripotent stem cells provided in Example 1 of the present invention;

[0058] Figure 3 A schematic diagram of calculating the Euclidean distance between two adjacent head pixel clusters to be detected in a method for detecting and identifying defective morphology of in vitro induced pluripotent stem cells provided in Example 1 of the present invention;

[0059] Figure 4 A flowchart of obtaining a circular area for a method for detecting and identifying defective morphology of in vitro induced pluripotent stem cells provided in Example 1 of the present invention;

[0060] Figure 5 A schematic diagram of vertical distance of a method for detecting, identifying and processing defective morphology of in vitro induced pluripotent stem cells provided in Example 1 of the present invention;

[0061] Figure 6 Defect information of the sperm head image to be detected is confirmed in the in vitro induced pluripotent stem cell defect morphology detection and identification processing method provided in the first embodiment of the present invention;

[0062] Figure 7 The shape and size of the sperm head deformity variation in the in vitro induced pluripotent stem cell defect morphology detection and identification processing method provided in Example 1 of the present invention;

[0063] Figure 8 A flowchart of a system for detecting and identifying defective morphology of in vitro induced pluripotent stem cells provided in Example 2 of the present invention;

[0064] Marking: image acquisition module 10, contour acquisition module 20, defect detection module 30. DETAILED DESCRIPTION

[0065] The technical solution of the present invention will be described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0066] The present invention is further described in detail below through specific embodiments in conjunction with the accompanying drawings.

[0067] Embodiment 1

[0068] like Figure 1 As shown, the present invention provides a method for detecting and identifying defective morphology of induced pluripotent stem cells in vitro, comprising the following specific steps:

[0069] S1: Obtain an image of the sperm head to be tested (i.e., a high-definition image of sperm cells obtained from in vitro induced pluripotent stem cells) and an image of the sample sperm head; before testing, first obtain a high-definition image of sperm cells differentiated from in vitro induced pluripotent stem cells. Due to the tiny size and complex structure of sperm cells, a high-resolution microscope is required to observe and capture images.

[0070] High-resolution microscopy can use phase contrast microscopy and fluorescence microscopy as well as electron microscopy; among them, phase contrast microscopy (Differential Interference Contrast Microscopy): Phase contrast microscopy is a commonly used microscopy technique that provides high-contrast and high-resolution images. It uses optical devices to enhance the details of the sample and produce images with a three-dimensional sense. Fluorescence Microscopy: Fluorescence microscopy uses specific molecules or structures labeled with fluorescent dyes to obtain images. By selecting appropriate fluorescent dyes to label specific components of sperm cells, the structure and distribution of sperm cells can be observed under a microscope.

[0071] Electron microscopy is a high-resolution microscopy technique that uses a beam of electrons rather than a light beam to view a sample. Transmission electron microscopy can provide very high resolution and is used to view the ultrastructure of sperm cells.

[0072] S2: Acquire the contours of the sperm head image to be detected and the sample sperm head image to obtain the contour of the head to be detected and the contour of the head sample; perform defect matching on the contour of the head to be detected and the contour of the head sample, output the defect matching result, and determine the defect information according to the defect matching result;

[0073] It should be noted that the sperm head image to be tested is an image of a sperm head that may have deformed or mutated characteristics, while the sample sperm head image is an image of a normal sperm head, and the normal sperm head image is an elliptical shape. The head head image to be tested and the sample sperm head image are used to obtain the head contour to be tested and the head sample contour, and by matching the areas of the head contour to be tested and the head sample contour, it is determined whether the sperm head image to be tested has a deformed or mutated defect, and finally the defect information (i.e., the defect result of the sperm head) is determined.

[0074] Specifically, Figure 2 As shown, in step S2, the contours of the sperm head image to be detected and the sample sperm head image are acquired to obtain the contour of the head to be detected and the contour of the head sample; defect matching is performed on the contour of the head to be detected and the contour of the head sample, and a defect matching result is output. Defect information is determined according to the defect matching result. The specific operation steps are as follows:

[0075] S21: preprocessing the sperm head image to be detected to obtain a grayscale image of the sperm to be detected;

[0076] Calculating grayscale values ​​of pixels of the grayscale image of sperm to be detected; clustering significant pixels in the grayscale image of sperm to be detected to obtain a plurality of head pixel clusters to be detected;

[0077] It should be noted that the pixels with high gray values ​​in the head image to be detected are clustered: because, according to the gray values ​​of the pixels of the sperm head and the background image of the head image to be detected, there will be changes in the gray values ​​of the sperm head and the background image in the head image to be detected, and the gray values ​​of the pixels of the sperm head will be higher than the gray values ​​of the pixels of the background image; therefore, the sperm head is a comparison of the gray values ​​of the pixels relative to the background image, and it often has pixels with high gray values ​​(i.e., significant pixels), so the pixels with high gray values ​​are clustered (i.e., significant pixel clustering), and several head pixel clusters to be detected can be obtained;

[0078] S22: Execute the judgment of whether the current head pixel cluster to be detected and the adjacent head pixel cluster to be detected belong to the same sperm cell: calculate the Euclidean distance between the head pixel cluster to be detected and the adjacent head pixel cluster to be detected, traverse the Euclidean distance between all two adjacent head pixel clusters to be detected; preset the Euclidean distance threshold w (i.e., the standard set Euclidean distance threshold w), and judge whether the Euclidean distance between all two adjacent head pixel clusters to be detected is less than the preset Euclidean distance threshold w;

[0079] If not, the head pixel cluster to be detected in the neighborhood is eliminated (then it is determined that the two do not belong to the same sperm cell); if so, it is determined that the current head pixel cluster to be detected and the adjacent head pixel cluster to be detected belong to the same sperm cell, and the current head pixel cluster to be detected and the adjacent head pixel cluster to be detected are obtained as the head target contour points to be detected, and all the confirmed head target contour points to be detected are connected to obtain the head contour to be detected; the closer the above Euclidean distance is, the more similar the two pixel clusters are, and the possibility that the two belong to the contour points of the same sperm cell is very high;

[0080] It should be noted that when the sperm head is deformed, there may be two sperm heads; therefore, the Euclidean distance between the two adjacent head pixel clusters to be detected is calculated, and it is determined whether the Euclidean distance between the two adjacent head pixel clusters to be detected is less than a preset Euclidean distance threshold w;

[0081] If it is greater than or equal to the preset Euclidean distance threshold w, it means that there may be cells belonging to two sperm heads. When there are cells belonging to two sperm heads, the head pixel clusters to be detected at the edges of the two sperm heads may be very close (that is, the head target contour points to be detected at the edges of the two sperm heads may be very close). When calculating the Euclidean distance between two adjacent head pixel clusters to be detected, if there are cells with two sperm heads, the Euclidean distances of the head pixel clusters to be detected at the edges of the two sperm heads may be calculated together (such as Figure 3 As shown), the Euclidean distance between the head pixel clusters to be detected at the edge of the two sperm heads is greater than or equal to the preset Euclidean distance threshold w. That is, it should be noted that if there are two sperm heads, the head pixel clusters to be detected at the edge are calculated together, and the target contour points of the head to be detected may be connected to the head contour to be detected at the end, and the head contour to be detected may have an erroneous connection contour. In this way, after confirming that the image of the sperm head to be detected has a deformed variation, it is impossible to correctly judge the shape of the specific deformed variation (that is, when the Euclidean distance of the wrong head pixel clusters to be detected is calculated, the head contours to be detected of the two sperm heads will be very large, and may even exceed the large-headed sperm, so when compared with the sample head contour in the subsequent test, it will definitely be detected as a deformed sperm head, but it is impossible to correctly judge what shape the deformed variation is, whether it is a large-headed sperm or a double-headed sperm);

[0082] If it is less than the preset Euclidean distance threshold w, it means that the two head-to-be-detected pixel clusters are the head-to-be-detected target contour points inside a sperm head and belong to a head-to-be-detected contour;

[0083] S23: preprocessing the sample sperm head image to obtain a sample sperm grayscale image;

[0084] Calculating grayscale values ​​of pixel points of the sample sperm grayscale image; clustering significant pixel points in the sample sperm grayscale image to obtain a plurality of head sample pixel clusters;

[0085] S24: traversing the pixel points inside the head sample pixel cluster to calculate the average grayscale value, and obtaining the average grayscale value of the head sample pixel cluster; calculating the grayscale mean of the average grayscale values ​​of all the head sample pixel clusters, assuming that the grayscale mean of the average grayscale values ​​of all the head sample pixel clusters is a threshold value t, and judging whether the average grayscale value of the head sample pixel cluster is greater than the grayscale mean of the average grayscale values ​​of all the head sample pixel clusters as the threshold value t;

[0086] If not, the head sample pixel cluster is eliminated;

[0087] If yes, the head sample pixel cluster is retained and confirmed as the head sample target contour point, and all the retained head sample target contour points are connected to obtain the head sample contour;

[0088] It should be noted that the sample sperm head image is a normal sperm head image. Therefore, after obtaining several head sample pixel clusters, in order to further make the contour of the normal sample sperm head image clearer, the grayscale mean of the average grayscale value of all head sample pixel clusters is taken (the average grayscale value of the internal pixel points of each head sample pixel cluster is calculated first, and then the grayscale mean of the average grayscale value of all head sample pixel clusters is calculated). The grayscale mean is used as the threshold t to screen out clear head sample target contour points, so that the final head sample contour is clearer.

[0089] S25: Determine a circular bounding box with a minimum radius by enclosing the contour of the head to be detected and the contour of the head sample, and obtain the circular area of ​​the head to be detected and the circular area of ​​the head sample; obtain the defect matching result of the sperm head image to be detected by comparing the circular area of ​​the head to be detected and the circular area of ​​the head sample; and determine the defect information according to the defect matching result;

[0090] It should be noted that since the shapes of the head contour to be detected and the head sample contour may be different, there may be errors when comparing the contours. Therefore, the area of ​​the head contour to be detected and the head sample contour is calculated for comparison; the defect matching result of the current sperm head image to be detected is judged by the difference between the two areas, and finally the defect information (i.e., defect type) is determined.

[0091] It needs to be further explained that, in the implementation of step S25, this embodiment still needs to further solve two problems. The first problem is the feasibility of area comparison. The researchers of this embodiment have discovered a fact that the implementation of the above steps is only suitable for area comparison and identification of mature sperm cells; the second problem is the irregularity of the head area.

[0092] Since sperm cells are also divided into different cycles of growth, maturity and senescence, their morphology is quite different; the image of the sperm head to be detected selected in step S25 of this embodiment is a mature sperm cell screened by a cell culture device, and the grayscale image of the same sample sperm is also an image of a mature sperm sample cell screened by the same cell culture device under the same culture environment; the reason for selecting the mature stage is that its morphology is relatively stable and has defect research value, and the error is large when identifying defects in the production period. However, during the implementation of this embodiment, there is still a further problem to be faced, that is, the head of the sperm cell is not a regular circle, especially the head shape of the deformed sperm cell is more irregular. For this, this embodiment uses the processing method of subsequent steps S251-S253 to screen the most contour pixels to determine a simulated circular shape representing the head morphology of the sperm, see steps S251-S253 for details.

[0093] Specifically, Figure 4 , 5 As shown, in step S25, the contour of the head to be detected and the contour of the head sample are bounded by a bounding box to determine a circular bounding box with a minimum radius, and the circular area of ​​the head to be detected and the circular area of ​​the head sample are obtained; by comparing the circular area of ​​the head to be detected and the circular area of ​​the head sample, the defect matching result of the sperm head image to be detected is obtained, and the specific operation steps are as follows:

[0094] S251: randomly selecting two edge pixel points on the head contour to be detected in a minimum enclosing circle manner, and connecting the two edge pixel points with a straight line to obtain a point-to-point straight line;

[0095] The two edge pixel points are two opposite points on the head contour to be detected, and after selecting the first edge pixel point, the edge pixel point on the head contour to be detected that is farthest from the first edge pixel point is selected as the second edge pixel point, and the first edge pixel point and the second edge pixel point are used as two opposite points;

[0096] It should be noted that two edge pixel points on the head contour to be detected are randomly selected by the minimum enclosing circle method to form a point-to-point straight line. The distribution of the remaining edge pixel points on the head contour to be detected can be determined by the point-to-point straight line. The edge pixel point closest to the point-to-point straight line is used as the circle inside the head contour to be detected.

[0097] S252: Calculate the vertical distances between the remaining edge pixel points on the head contour to be detected and the point-to-point straight line, preset a vertical distance threshold j, and determine whether the current vertical distance is less than the preset vertical distance threshold j;

[0098] If so, retain the edge pixel points for calculating the vertical distance;

[0099] If not, the edge pixels for calculating the vertical distance are eliminated;

[0100] Traverse all the remaining edge pixel points on the head contour to be detected to see if the vertical distance to the point straight line is greater than the preset vertical distance threshold j;

[0101] Selecting two edge pixel points with the farthest distance from each other from the retained edge pixel points as diameters to form a minimum bounding box, and calculating the area of ​​the minimum bounding box as the circular area of ​​the head to be detected;

[0102] It should be noted that by calculating the vertical distance between the remaining edge pixels and the straight line to determine whether it is greater than the preset vertical distance threshold j, it can be determined whether these edge pixels have an impact on the head contour to be detected. Because the head contour to be detected is a curved contour shape, when the vertical distance is less than or equal to the preset vertical distance threshold j, the edge pixel point will deviate from the straight line and cannot form a vertical distance with the straight line; and the position of the edge pixel point on the head contour to be detected is the position of the irregular curve (that is, on the irregular contour of the head contour to be detected). When the edge pixel point at this position is looking for the edge pixel point with the farthest distance, the minimum bounding box will not be the smallest circle;

[0103] When the vertical distance is greater than the preset vertical distance threshold j, it means that the edge pixel point is relatively close to the point-to-point straight line, which has a small impact on the curved contour shape of the head contour to be detected, and when the minimum bounding box is determined by the retained edge pixel points, a tighter and more accurate minimum bounding box can be obtained; because the minimum bounding box is determined based on the two farthest retained edge pixel points on the head contour to be detected, and the two retained edge pixel points are far away from the point-to-point straight line, which helps to ensure that the minimum bounding box can fit the head contour closely;

[0104] In summary, all edge pixels are circled in a minimum enclosing circle manner to determine a minimum enclosing box (i.e., a circular enclosing box with a minimum radius); the radius and area of ​​the minimum enclosing box are calculated; and the final area of ​​the current minimum enclosing box is regarded as: the circular area of ​​the head to be detected.

[0105] S253: Execute steps S251 and S252 on the head sample contour to obtain the circular area of ​​the head sample;

[0106] S254: Match the circular area of ​​the head to be detected with the circular area of ​​the head sample, and calculate the circular area difference; preset a circular area threshold u, and determine whether the circular area difference is less than the preset circular area threshold u; it should be noted that the defect matching result of the sperm head image to be detected specifically refers to the area difference comparison result.

[0107] If not, it is determined that the head image to be detected of the current sperm head image to be detected has defects, and the specific defect information is confirmed through the defect matching result of the sperm head image to be detected;

[0108] If yes, then the head image to be detected of the sperm head image to be detected has no deformity variation;

[0109] It should be noted that, because the sperm head is elliptical, when calculating the contour area, calculation errors or excessive calculation data may occur, resulting in a decrease in matching efficiency. Therefore, the area of ​​the circle is calculated by the minimum bounding box method, and finally the matching result of the head image to be detected of the sperm head image to be detected is determined according to the size difference between the circular area of ​​the head to be detected and the circular area of ​​the head sample; when the area difference between the two is smaller, it means that the head image to be detected of the sperm head image to be detected is more normal, and there is no deformity or variation in the sperm head; when the area difference between the two is larger, it means that the head image to be detected of the sperm head image to be detected is more abnormal, and there is a defect of deformed shape, and the specific defect information (i.e., defect type) is specifically confirmed through the deformed shape. For example, the sperm head may be a large-headed sperm or a micro sperm (i.e., a small-headed sperm), a pear-shaped sperm, or a double-headed sperm, which are four typical head defects;

[0110] And by subsequently determining what deformed shape the sperm head image to be tested has, it is possible to further confirm whether the sperm head image to be tested actually has the deformed defect, so that it can be accurately confirmed whether the sperm head image to be tested has the deformed defect.

[0111] Specifically, Figure 6 As shown, in step S254, the specific defect information is confirmed by the defect matching result of the sperm head image to be detected, and the specific operation steps are as follows:

[0112] By comparing the preset circular area threshold u with the circular area difference, it is determined that the image of the sperm head to be detected has deformed variation, but it is impossible to confirm the specific shape of the deformed variation (i.e., the sperm head may be a large-headed sperm or a micro sperm (i.e., a small-headed sperm), a pear-shaped sperm, or a double-headed sperm);

[0113] And after research, it was found that among the shapes of sperm head deformities, each shape is different in size, and the sizes are as follows: micro sperm (i.e. small-headed sperm) < pear-shaped sperm < normal sperm head < double-headed sperm < large-headed sperm, such as Figure 7 As shown;

[0114] Therefore, when it is determined that there is a deformity variation in the sperm head image to be detected by the preset circular area threshold u, the shape and size of the sperm head deformity variation (i.e., the size of the area) are used to confirm what shape the deformity variation is, and further operations are performed based on the circular area difference between the circular area of ​​the head to be detected and the circular area of ​​the head sample. The specific operation steps are as follows:

[0115] S2541: when it is determined that the head image to be detected of the sperm head image to be detected has defects, a first defect threshold g and a second defect threshold h are preset, and a magnitude relationship between the circular area difference value in the first defect threshold g and the second defect threshold h is determined;

[0116] The first defect threshold g<the second defect threshold h;

[0117] S2542: If the circular area difference is less than the first defect threshold g, the defect information of the sperm head image to be detected is micro sperm (i.e., small-headed sperm);

[0118] S2543: If the circular area difference = the first defect threshold g, the defect information of the sperm head image to be detected is pear-shaped sperm;

[0119] S2544: If the circular area difference is greater than the second defect threshold h, the defect information of the sperm head image to be detected is a large-headed sperm;

[0120] S2545: If the circular area difference = the second defect threshold h, the defect information of the sperm head image to be detected is a double-headed sperm;

[0121] S2546: If the first defect threshold g is less than the circular area difference and less than the second defect threshold h, then the sperm head image to be detected has no defect problem;

[0122] It should be noted that, because the size and shape of the sperm head (i.e., the size of the area) are ranked as follows: micro sperm (i.e., small-headed sperm) < pear-shaped sperm < normal sperm head < double-headed sperm < large-headed sperm; therefore, a double threshold is set to determine the range of the circular area difference;

[0123] Since micro sperm (i.e., small-headed sperm) has the smallest shape, when the circular area difference is less than the first defect threshold g, the defect information of the sperm head image to be detected is determined to be micro sperm (i.e., small-headed sperm);

[0124] And micro sperm (i.e. small-headed sperm) is smaller than pear-shaped sperm, and pear-shaped sperm belongs to the second-to-last small sperm head, so when the circular area difference is equal to the first defect threshold g, the defect information of the sperm head image to be detected is pear-shaped sperm;

[0125] At the same time, the big-headed sperm is the largest deformed variant shape. Therefore, when the circular area difference is greater than the second defect threshold h, the defect information of the sperm head image to be detected is the big-headed sperm.

[0126] Since the shape of a double-headed sperm is smaller than that of a large-headed sperm, when the circular area difference is equal to the second defect threshold h, the defect information of the sperm head image to be detected is a double-headed sperm;

[0127] If the first defect threshold g is less than the circular area difference and the second defect threshold h is less than the circular area difference, it means that the circular area difference is within the double threshold range, and the interior of the sperm head image to be detected is a normal sperm head image;

[0128] Through the judgment of double thresholds, not only can the deformed shape of the sperm head image to be tested be determined, but also the double thresholds can be used to further confirm whether the sperm head image to be tested really has the defect of deformed variation, making the result more accurate.

[0129] Embodiment 2

[0130] like Figure 8 As shown, accordingly, the present invention also proposes an in vitro induced pluripotent stem cell defect morphology detection and recognition processing system, comprising: an image acquisition module 10, a contour acquisition module 20, and a defect detection module 30;

[0131] The image acquisition module 10 is used to acquire the image of the sperm head to be detected and the image of the sample sperm head;

[0132] The contour acquisition module 20 is used to acquire contours of the sperm head image to be detected and the sample sperm head image to obtain the head contour to be detected and the head sample contour;

[0133] The defect detection module 30 is used to perform defect matching on the head contour to be detected and the head sample contour, output a defect matching result, and determine defect information according to the defect matching result.

[0134] In summary, an in vitro induced pluripotent stem cell defect morphology detection and identification processing method and system proposed in an example of the present invention converts the sperm head image to be detected into a sperm grayscale image to be detected, obtains the sperm grayscale image to be detected, and calculates the grayscale value of the pixel point. The grayscale value of the pixel point of the sperm head inside the sperm grayscale image to be detected is different from the grayscale value of the pixel point of the background image of the sperm grayscale image to be detected. The grayscale value of the pixel point of the sperm head will be very high, so the pixel points with high grayscale values ​​are clustered to obtain a number of head pixel clusters to be detected;

[0135] Calculate the Euclidean distance between the head pixel clusters to be detected, and determine whether the head pixel clusters to be detected belong to one sperm head cell or possibly belong to two sperm head cells; confirm the head pixel clusters to be detected that meet the conditions as the head target contour points to be detected and connect them to obtain the head target contour points to be detected; the sample sperm head image is a complete image, so the pixel points with high grayscale values ​​of the sample sperm head image are clustered to obtain the head sample pixel clusters; then use the threshold to screen the pixel points inside the head sample pixel cluster again to select the head sample pixel cluster with high average grayscale value, and confirm it as the head sample target contour point, so that clear head sample target contour points are screened out twice, and a clear head sample contour is connected through the clear head sample target contour points;

[0136] Further, two random edge pixel points of the head contour to be detected are selected by the minimum enclosing circle method, and the two edge pixel points are connected to obtain a point-to-point straight line; the vertical distances between the remaining edge pixel points and the point-to-point straight line are determined, and the non-conforming edge pixel points are eliminated, and the retained edge pixel points are used to further screen out the two conforming edge pixel points to form a diameter, and the minimum enclosing box is obtained by the diameter, and the circular area of ​​the head to be detected is further obtained; the circular area of ​​the head sample contour is obtained through the above steps;

[0137] Furthermore, the circular area of ​​the head to be detected is compared with the circular area of ​​the head sample, and the circular area difference is used to determine whether the image of the sperm head to be detected has defects; if the circular area difference is smaller than the preset circular area threshold, it means that the circular area of ​​the head to be detected and the circular area of ​​the head sample are close to each other, and if they are close to each other, the image of the sperm head to be detected has no defects; if it is larger than the preset circular area threshold, it means that the image of the sperm head to be detected has defects, and the defect information of the sperm head image to be detected is determined by the defects.

[0138] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. A person skilled in the art may modify the technical solutions described in the above embodiments, or replace part or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for detecting and identifying defective morphology of induced pluripotent stem cells in vitro, characterized in that: Including specific steps: Acquire the image of the sperm head to be detected and the image of the sample sperm head; Acquiring the contour of the sperm head image to be detected to obtain the contour of the head to be detected; Acquiring the contour of the sample sperm head image to obtain a head sample contour; Perform defect matching on the head contour to be detected and the head sample contour, output a defect matching result, and determine defect information according to the defect matching result; The contour of the sperm head to be detected is obtained by acquiring the contour of the head to be detected. The specific operation steps are as follows: Preprocessing the sperm head image to be detected to obtain a sperm grayscale image to be detected; Calculating grayscale values ​​of pixels of the grayscale image of sperm to be detected; clustering significant pixels in the grayscale image of sperm to be detected to obtain a plurality of head pixel clusters to be detected; Execute the determination of whether the current head pixel cluster to be detected and the adjacent head pixel cluster to be detected belong to the same sperm cell: calculate the Euclidean distance between the head pixel cluster to be detected and the adjacent head pixel cluster to be detected, and traverse the Euclidean distances between all two adjacent head pixel clusters to be detected; A Euclidean distance threshold w is preset, and it is determined whether the Euclidean distance between all two adjacent head pixel clusters to be detected is less than the preset Euclidean distance threshold w; If not, the head pixel cluster to be detected in the neighborhood is eliminated; If so, it is determined that the current head pixel cluster to be detected and the adjacent head pixel cluster to be detected belong to the same sperm cell, the current head pixel cluster to be detected and the adjacent head pixel cluster to be detected are obtained as the head target contour points to be detected, and all the confirmed head target contour points to be detected are connected to obtain the head contour to be detected; Defect matching is performed on the head contour to be detected and the head sample contour, a defect matching result is output, and defect information is determined according to the defect matching result. The specific operation steps are as follows: The contour of the head to be detected and the contour of the head sample are bounded by a bounding box to determine a circular bounding box with a minimum radius, and the circular area of ​​the head to be detected and the circular area of ​​the head sample are obtained; by comparing the circular area of ​​the head to be detected and the circular area of ​​the head sample, a defect matching result of the sperm head image to be detected is obtained; and defect information is determined according to the defect matching result; Perform bounding box demarcation on the head contour to be detected and the head sample contour to determine a circular bounding box with a minimum radius, and obtain a circular area of ​​the head to be detected and a circular area of ​​the head sample; Randomly select two edge pixel points on the head contour to be detected in a minimum enclosing circle manner, connect the two edge pixel points with a straight line, and obtain a point-to-point straight line; The two edge pixel points are two opposite points on the head contour to be detected, and after selecting the first edge pixel point, the edge pixel point on the head contour to be detected that is farthest from the first edge pixel point is selected as the second edge pixel point, and the first edge pixel point and the second edge pixel point are used as two opposite points; Calculate the vertical distance between the remaining edge pixel points on the head contour to be detected and the point straight line, preset a vertical distance threshold j, and determine whether the current vertical distance is greater than the preset vertical distance threshold j; If so, retain the edge pixel points for calculating the vertical distance; If not, the edge pixels for calculating the vertical distance are eliminated; Traverse all the remaining edge pixel points on the head contour to be detected to see if the vertical distance to the point straight line is greater than the preset vertical distance threshold j; Selecting two edge pixel points with the farthest distance from each other from the retained edge pixel points as diameters to form a minimum bounding box, and calculating the area of ​​the minimum bounding box as the circular area of ​​the head to be detected; The above steps are performed on the head sample contour to obtain the head sample circular area.

2. According to claim 1, a method for detecting and identifying defective morphology of induced pluripotent stem cells in vitro, characterized in that: The contour of the sample sperm head image is obtained to obtain the head sample contour. The specific operation steps are as follows: Preprocessing the sample sperm head image to obtain a sample sperm grayscale image; Calculating grayscale values ​​of pixel points of the sample sperm grayscale image; clustering significant pixel points in the sample sperm grayscale image to obtain a plurality of head sample pixel clusters; Traversing the pixel points inside the head sample pixel cluster to calculate the average grayscale value, to obtain the average grayscale value of the head sample pixel cluster; Calculate the grayscale mean of the average grayscale values ​​of all head sample pixel clusters, set the grayscale mean of the average grayscale values ​​of all head sample pixel clusters as a threshold value t, and determine whether the average grayscale value of the head sample pixel cluster is greater than the grayscale mean of the average grayscale values ​​of all head sample pixel clusters as the threshold value t; If not, the head sample pixel cluster is eliminated; If so, the head sample pixel cluster is retained and confirmed as the head sample target contour point, and all the retained head sample target contour points are connected to obtain the head sample contour.

3. The method for detecting and identifying defective morphology of induced pluripotent stem cells in vitro according to claim 2, characterized in that: By comparing the circular area of ​​the head to be detected with the circular area of ​​the head sample, the defect matching result of the sperm head image to be detected is obtained. The specific operation steps are as follows: Match the circular area of ​​the head to be detected with the circular area of ​​the head sample, and calculate the circular area difference; preset a circular area threshold u, and determine whether the circular area difference is less than the preset circular area threshold u; If not, it is determined that the head image to be detected of the current sperm head image to be detected has defects, and the specific defect information is confirmed through the defect matching result of the sperm head image to be detected; If so, the head image to be detected of the sperm head image to be detected has no deformity or variation.

4. The method for detecting and identifying defective morphology of induced pluripotent stem cells in vitro according to claim 3, characterized in that: The specific defect information is confirmed by the defect matching result of the sperm head image to be detected, and the specific operation steps are as follows: When it is determined that the head image to be detected of the sperm head image to be detected has defects, a first defect threshold g and a second defect threshold h are preset, and a magnitude relationship between the circular area difference value in the first defect threshold g and the second defect threshold h is determined; The first defect threshold g<the second defect threshold h; If the circular area difference is less than the first defect threshold g, the defect information of the sperm head image to be detected is micro sperm; If the circular area difference = the first defect threshold g, then the defect information of the sperm head image to be detected is pear-shaped sperm; If the circular area difference is greater than the second defect threshold h, the defect information of the sperm head image to be detected is a large-headed sperm; If the circular area difference = the second defect threshold h, then the defect information of the sperm head image to be detected is a double-headed sperm; If the first defect threshold g<circular area difference<second defect threshold h, then the sperm head image to be detected has no defect problem.

Images