Stem cell culture image recognition and detection method

Abstract

The invention discloses a stem cell culture image identification and detection method, and relates to the technical field of stem cell image identification, and the method comprises the following steps: carrying out the graying processing of a stem cell image, obtaining a stem cell gray-scale image, carrying out the binarization processing of the stem cell gray-scale image, obtaining a stem cell binarization image, and obtaining a real-time complete contour based on the stem cell binarization image; acquiring a real-time length-width ratio; obtaining a range of a historical length-width ratio based on the first number of historical mesenchymal stem cells; the qualified real-time length-width ratio is obtained based on the range of the real-time length-width ratio and the historical length-width ratio, and the method is used for solving the problems that in an existing mesenchymal stem cell number recognition technology, the effective number of mesenchymal stem cells in the culture process cannot be accurately extracted, and consequently the accuracy and efficiency of recognizing the number of the mesenchymal stem cells are low; under the condition, the culture quality of the mesenchymal stem cells cannot be effectively judged.

Description

Technical Field

[0001] The present invention relates to the technical field of stem cell image recognition, and in particular to a stem cell culture image recognition detection method. Background Art

[0002] Mesenchymal stem cells are often cultured because of their multidirectional differentiation potential, self-renewal ability, immune regulation function and potential application in the treatment of various diseases. During the culture process of mesenchymal stem cells, the experimenters need to constantly observe the mesenchymal stem cells, the number of mesenchymal stem cells and whether they contain other cells, so as to facilitate timely corresponding treatment and operation; The existing method of observing the number of mesenchymal stem cells uses manual observation, which has many disadvantages, including poor accuracy and repeatability, susceptibility to subjective factors and fatigue, inability to work continuously for a long time, limited ability to recognize complex images, which may lead to data deviation and insufficient generalization ability, and inability to handle large amounts of data. These limitations limit its potential in terms of accuracy and efficiency, making automated image recognition technology an important supplement to improve the quality and efficiency of stem cell culture. In the existing technology of automated cell image recognition, there are deficiencies in the image feature extraction of stem cells, which cannot be applied to the number extraction of stem cells. For example, in the application document with publication number: CN104850860A, a cell image recognition method and a cell image recognition device are disclosed. This method reduces the amount of calculation by adopting the application of compressed sensing principle technology in cell recognition, but it lacks a method for identifying the characteristics of mesenchymal stem cells, and cannot accurately extract the effective number of mesenchymal stem cells in the culture process, resulting in low accuracy and efficiency in identifying the number of mesenchymal stem cells. In this case, the quality of mesenchymal stem cell culture cannot be effectively judged. Summary of the invention

[0003] The present invention aims to solve at least one of the technical problems in the prior art to a certain extent, by performing image processing on a stem cell image to obtain a real-time complete outline; obtaining a real-time rectangular outline of the real-time complete outline based on a complete outline frame selection method, and obtaining a real-time aspect ratio based on the real-time rectangular outline; obtaining a range of historical aspect ratios based on a first number of historical mesenchymal stem cells; obtaining a qualified real-time aspect ratio based on the real-time aspect ratio and the range of historical aspect ratios, and obtaining the number of qualified mesenchymal stem cells based on the number of all qualified real-time aspect ratios, so as to solve the problem that the existing mesenchymal stem cell number identification technology cannot accurately extract the effective number of mesenchymal stem cells in the culture process, resulting in low accuracy and efficiency in identifying the number of mesenchymal stem cells, and in this case, the quality of mesenchymal stem cell culture cannot be effectively judged.

[0004] To achieve the above objectives, the present application provides a stem cell culture image recognition detection method, comprising the following steps: The culture images of mesenchymal stem cells are obtained based on the microscope and are marked as stem cell images; Gray-processing the stem cell image to obtain a stem cell grayscale image, binarizing the stem cell grayscale image to obtain a stem cell binary image, and obtaining a real-time complete contour based on the stem cell binary image; Based on the complete outline selection method, a real-time rectangular outline of the real-time complete outline is obtained, and based on the real-time rectangular outline, a real-time aspect ratio is obtained; Obtaining a range of historical aspect ratios based on the first number of historical mesenchymal stem cells; A qualified real-time aspect ratio is obtained based on the range of the real-time aspect ratio and the historical aspect ratio, and the number of qualified mesenchymal stem cells is obtained based on the number of all qualified real-time aspect ratios.

[0005] Furthermore, grayscale processing of the stem cell image to obtain a stem cell grayscale image includes the following sub-steps: Obtain the R, G and B values ​​of the RGB value of each pixel of the stem cell image; Use the grayscale conversion formula to convert the RGB value into a grayscale value. The grayscale conversion formula is: Hd=p1*R+p2*G+p3*B; Hd is the grayscale value of the pixel in the stem cell image, and p1, p2 and p3 are the R, G and B weights respectively.

[0006] Furthermore, binarizing the stem cell grayscale image to obtain a stem cell binarization image includes the following sub-steps: Divide the grayscale value of 0-255 into f intervals evenly, marked as grayscale intervals; Count the frequency of the grayscale values ​​in each grayscale interval respectively, and mark it as grayscale interval frequency; Draw a histogram with grayscale value as X-axis and grayscale interval frequency as Y-axis, marked as grayscale histogram; Add up the frequencies of all grayscale intervals to get the total frequency, marked as J; Mark the intervals with grayscale interval frequency less than J / f as smaller areas, and mark the intervals other than the smaller areas as larger areas. Determine whether there are larger areas on both sides of the smaller area. If so, mark the smaller area as the final area. The midpoint value of the abscissa of the final area is set as the background threshold; the grayscale value of the pixel points greater than or equal to the background threshold is set to 255, and the grayscale value of the pixel points less than the background threshold is set to 0.

[0007] Furthermore, obtaining a real-time complete outline based on the stem cell binary image includes the following sub-steps: If a pixel with a gray value of 255 is adjacent to a pixel with a gray value of 0, the pixel with a gray value of 0 is marked as a stem cell edge pixel; All pixels except those at the edge of stem cells were set to 255; Obtain any pixel point at the edge of a stem cell, mark it as the initial center pixel point, establish a 3x3 pixel area with the initial center pixel point as the center, mark it as the reference area, and the nine pixels in the reference area include the initial center pixel point and eight neighboring pixels around the initial center pixel point; Determine whether the eight neighboring pixels around the initial center pixel point contain a stem cell edge pixel point. If so, mark the stem cell edge pixel point adjacent to the initial center pixel point as an adjacent edge pixel point. If not, remove the initial center pixel point and continue to select the next arbitrary stem cell edge pixel point as the initial center pixel point. A new reference area is established again by taking any adjacent edge pixel point as the center pixel point of the reference area, obtaining the newly appeared stem cell edge pixel point from the new reference area and marking it as a new edge pixel point, and the new edge pixel point is used as the center pixel point of the new reference area to establish a new reference area again, and so on, and new reference areas are continuously established until there are no new stem cell edge pixels in the new reference area; Determine whether the pixels in the reference area finally established contain the initial central pixel. If so, obtain the central pixel of all reference areas and mark them as completed contours. If it does not contain the initial central pixel, obtain the central pixel of the last established reference area, mark it as the final central pixel, establish a 5x5 pixel area with the final central pixel as the center, mark it as the final reference area, and determine whether the final reference area contains the initial central pixel. If it contains the initial central pixel, obtain the pixels on the straight path between the initial central pixel and the final central pixel, mark them as newly added contour pixels, set the grayscale value of the newly added contour pixels to 0, obtain the central pixels and newly added contour pixels of all reference areas, and mark them as completed contours; if it does not contain the initial central pixel, obtain the central pixels of all reference areas, mark them as real-time incomplete contours, and refer to completed contours and completed contours as real-time complete contours; remove real-time incomplete contours.

[0008] Further, obtaining a real-time rectangular outline of a real-time complete outline based on the complete outline selection method, and obtaining a real-time aspect ratio based on the real-time rectangular outline includes the following sub-steps: The complete outline selection method includes: Establishing a real-time plane rectangular coordinate system, placing each individual real-time complete profile in the first quadrant; Set the coordinate point of the real-time complete contour in the real-time rectangular coordinate system to (x i ,y i ), get an x i The maximum coordinates of max , y1), get an x i The minimum value of x is: min , y2), get a y i The maximum coordinates are: (x1, y max ), get y i The coordinates of the minimum value are: (x2, y min Then get (x max , y1) and perpendicular to the X-axis, through (x min , y2) and perpendicular to the X-axis, through (x1, y max ) and perpendicular to the Y axis and the line segment passing through (x2, y min ) and perpendicular to the Y-axis, the four line segments form a rectangle; Mark the rectangle obtained by the real-time complete outline using the complete outline frame selection method as the real-time rectangle outline; Get the length and width of the real-time rectangular outline, marked as Cs and Ks respectively; The real-time aspect ratio is calculated as: Qs=Cs / Ks; where Qs is the real-time aspect ratio.

[0009] Further, obtaining the range of the historical aspect ratio based on the first number of historical mesenchymal stem cells includes the following sub-steps: Obtain the first quantity of historical mesenchymal stem cells, obtain the complete historical profile of historical mesenchymal stem cells; For any historical complete contour, the historical complete contour is obtained into a rectangle by using the complete contour box selection method, and is marked as the historical rectangular contour; Get the length and width of the historical rectangular outline, marked as Cl and Kl respectively; The historical aspect ratio is calculated as: Ql=Cl / Kl; where Ql is the historical aspect ratio.

[0010] Furthermore, obtaining the range of the historical aspect ratio based on the first number of historical mesenchymal stem cells further includes the following sub-steps: Evenly divide the historical aspect ratio into multiple intervals, marked as historical aspect ratio intervals, and count the frequency of the historical aspect ratio in each aspect ratio interval, marked as interval historical aspect ratio frequency; With the historical aspect ratio as the X-axis and the interval historical aspect ratio frequency as the Y-axis, a histogram is drawn and marked as the historical aspect ratio histogram.

[0011] Further, the steps for obtaining the range of historical aspect ratios based on the first quantity of historical mesenchymal stem cells further include the following sub-steps: Count the number of historical aspect ratio intervals, denoted as Z1; Sort the historical aspect ratio frequencies of the intervals in ascending order, and assign an ordinal number to each historical aspect ratio frequency of the interval. The ordinal number is an integer starting from 1; Determine whether w1 * Z1 is an integer. If it is an integer, set the historical aspect ratio frequency of the interval with the serial number w1 * Z1 as the first ratio number, denoted as T1; if it is not an integer, calculate the average value of the historical aspect ratio frequencies of the intervals corresponding to the serial numbers on both sides of w1 * Z1 as the first ratio number; where w1 is the first ratio; Determine whether w2 * Z1 is an integer. If it is an integer, set the historical aspect ratio frequency of the interval with the serial number w2 * Z1 as the second quantile, denoted as T2; if it is not an integer, calculate the average value of the historical aspect ratio frequencies of the intervals corresponding to the serial numbers on both sides of w2 * Z1 as the second quantile; where w2 is the first ratio, and w2 = [(1 / w1) - 1] * w1; Set the过小 threshold as: t1 = T1 - r * (T2 - T1); where t1 is the过小 threshold and r is the range constant.

[0012] Further, the steps for obtaining the range of historical aspect ratios based on the first quantity of historical mesenchymal stem cells further include the following sub-steps: Determine whether all the historical aspect ratio frequencies of the intervals are less than the过小 threshold. If so, mark the historical aspect ratio intervals corresponding to the historical aspect ratio frequencies less than the过小 threshold as the过小区间, delete the过小区间 in the historical aspect ratio histogram, and mark the historical aspect ratio histogram after deleting the过小区间 as the final aspect ratio histogram; if not, directly mark the historical aspect ratio histogram as the final aspect ratio histogram; Obtain the minimum value and the maximum value of the historical aspect ratio intervals in the X-axis data of the final aspect ratio histogram, denoted as Qmin and Qmax respectively.

[0013] Further, obtaining the qualified real-time aspect ratio based on the range of the real-time aspect ratio and the historical aspect ratio, and obtaining the number of qualified mesenchymal stem cells based on the number of all qualified real-time aspect ratios include the following sub-steps: Determine whether all the real-time aspect ratios satisfy Qmin < Qs < Qmax. If so, retain the real-time rectangular contour; if not, delete the real-time rectangular contour; obtain the number of all real-time rectangular contours, denoted as the number of qualified mesenchymal stem cells.

[0014] Beneficial effects of the present invention: The present invention obtains a real-time complete outline by performing image processing on a stem cell image; obtains a real-time rectangular outline of the real-time complete outline based on a complete outline frame selection method, and obtains a real-time aspect ratio based on the real-time rectangular outline; obtains a range of historical aspect ratios based on a first number of historical mesenchymal stem cells; obtains a qualified real-time aspect ratio based on the real-time aspect ratio and the range of historical aspect ratios, and obtains the number of qualified mesenchymal stem cells based on the number of all qualified real-time aspect ratios. The advantage of the present invention is that the number of qualified mesenchymal stem cells is obtained by using image processing and data processing, thereby improving the accuracy and processing speed of stem cell identification; The present invention obtains a real-time complete contour by using a reference area. The advantage is that all stem cell edge pixels can be screened out by using the reference area. Based on the stem cell edge pixels searched in the reference area, a real-time complete contour and a real-time incomplete contour are obtained. The real-time incomplete contour is eliminated, so that the subsequent judgment of whether it is a mesenchymal stem cell is more accurate, thereby improving the accuracy of mesenchymal stem cell quantity extraction. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 is a functional block diagram of the system of the present invention; Figure 2 is a schematic diagram of a grayscale histogram of the present invention; Figure 3 A schematic diagram of various reference areas established with an initial central pixel point according to the present invention; Figure 4 A schematic diagram of the process of continuously establishing reference regions of the present invention; Figure 5 A schematic diagram of the process of obtaining newly added contour pixels of the present invention; Figure 6 A schematic diagram of obtaining a real-time rectangular outline of the present invention; Figure 7 Schematic diagram of the historical aspect ratio histogram of the present invention. DETAILED DESCRIPTION

[0016] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. 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.

[0017] Example 1, please refer to Figure 1As shown, a stem cell culture image recognition and detection method includes: step S1, obtaining a culture image of mesenchymal stem cells based on a microscope and marking it as a stem cell image; in order to facilitate the identification of mesenchymal stem cells during the stem cell culture process, it is usually necessary to stain the mesenchymal stem cells, so the obtained culture image of the mesenchymal stem cells is a staining acquisition image.

[0018] Step S2, see Figure 2 As shown, the stem cell image is grayed to obtain a stem cell grayscale image, the stem cell grayscale image is binarized to obtain a stem cell binarized image, and a real-time complete contour is obtained based on the stem cell binarized image; step S2 includes the following sub-steps: Step S201, obtaining the R, G and B values ​​of the RGB value of each pixel of the stem cell image; converting the RGB value into a grayscale value using a grayscale conversion formula, the grayscale conversion formula is: Hd=p1*R+p2*G+p3*B; Hd is the gray value of the pixel in the stem cell image, p1, p2 and p3 are the weights of R, G and B respectively; the weighted average method is used to determine the weights of the three color channels of red, green and blue according to the sensitivity of the human eye to different colors, and then the RGB value of each pixel is weighted averaged to obtain the corresponding gray value, such as p1, p2 and p3 are 0.299, 0.587 and 0.114 respectively; Step S202, divide the grayscale values ​​of 0-255 into f intervals on average, marked as grayscale intervals; count the frequency of the grayscale values ​​in each grayscale interval respectively, marked as grayscale interval frequency; f can be set according to a specific method, such as being set to 8; Step S203, draw a histogram with the gray value as the X-axis and the gray interval frequency as the Y-axis, marked as a gray histogram; Step S204, add all the grayscale interval frequencies to get the total frequency, marked as J; mark the intervals with grayscale interval frequencies less than J / f as smaller areas, mark the intervals other than the smaller areas as larger areas, determine whether there are larger areas on both sides of the smaller area, if so, mark the smaller area as the final area; set the midpoint value of the abscissa of the final area as the background threshold; set the grayscale value of the pixel points greater than or equal to the background threshold to 255, and the grayscale value of the pixel points less than the background threshold to 0. Because active mesenchymal stem cells are usually stained red in culture, the converted grayscale values ​​are generally distributed around 70, while the background is generally white or colorless, and the converted grayscale values ​​are distributed in a larger range, the distribution of the grayscale values ​​should be different, so the background threshold can be obtained by observing the histogram; Step S205, if a pixel with a gray value of 255 is adjacent to a pixel with a gray value of 0, the pixel with a gray value of 0 is marked as a stem cell edge pixel; all pixels except the stem cell edge pixel are set to 255; only the gray value of the stem cell edge pixel is kept to 0, and the gray values ​​of the remaining pixels are set to 255; Step S206, obtaining any stem cell edge pixel point, marking it as the initial center pixel point, establishing a 3x3 pixel area with the initial center pixel point as the center, marking it as the reference area, the nine pixel points in the reference area include the initial center pixel point and the eight neighboring pixel points around the initial center pixel point; the principle is: the complete stem cell contour is closed, so the complete stem cell contour stem cell edge pixel points all have adjacent stem cell contour stem cell edge pixel points, searching all stem cell contour stem cell edge pixel points until the initial center pixel point is the complete contour; Step S207, see Figure 3 and Figure 4 As shown, it is determined whether the eight neighborhood pixels around the initial center pixel contain stem cell edge pixels. If so, the stem cell edge pixels adjacent to the initial center pixel are marked as adjacent edge pixels. If not, the initial center pixel is removed, and the next arbitrary stem cell edge pixel is selected as the initial center pixel; any adjacent edge pixel is selected as the center pixel of the reference area to establish a new reference area again, and the newly appeared stem cell edge pixel is obtained from the new reference area and marked as a new edge pixel. The new edge pixel is used as the center pixel of the new reference area to establish a new reference area again, and so on, and new reference areas are continuously established until there are no new stem cell edge pixels in the new reference area; it is determined whether the eight neighborhood pixels around the initial center pixel contain stem cell edge pixels, because there may be independent initial center pixels; Step S208, see Figure 5As shown, it is determined whether the pixels in the last established reference area contain the initial central pixel. If the initial central pixel is contained, the central pixels of all the reference areas are obtained and marked as the completed contour. If the initial central pixel is not contained, the central pixel of the last established reference area is obtained and marked as the final central pixel. A 5x5 pixel area is established with the final central pixel as the center and marked as the final reference area. It is determined whether the final reference area contains the initial central pixel. If the initial central pixel is contained, the pixels on the straight path between the initial central pixel and the final central pixel are obtained and marked as the newly added contour pixels. The gray value of the newly added contour pixels is set to 0, and the central pixels of all the reference areas and the newly added contour pixels are obtained and marked as the completed contour. If the initial central pixel is not contained, the central pixels of all the reference areas are obtained and marked as the real-time incomplete contour. The completed contour and the completed complete contour are collectively referred to as the real-time complete contour. The real-time incomplete contour is eliminated. The final central pixel is established because there may be a contour that is closed by only one or two pixels. Such a contour can be regarded as a complete contour.

[0019] Step S3, obtaining a real-time rectangular outline of a real-time complete outline based on the complete outline selection method, and obtaining a real-time aspect ratio based on the real-time rectangular outline; Step S3 includes the following sub-steps: Step S301, the complete contour selection method includes: establishing a real-time plane rectangular coordinate system, placing each individual real-time complete contour in the first quadrant; setting the coordinate point of the real-time complete contour in the real-time rectangular coordinate system as (x i ,y i ), get an x i The maximum coordinates of max , y1), get an x i The minimum value of x is: min , y2), get a y i The maximum coordinates are: (x1, y max ), get y i The coordinates of the minimum value are: (x2, y min Then get (x max , y1) and perpendicular to the X-axis, through (x min , y2) and perpendicular to the X-axis, through (x1, y max ) and perpendicular to the Y axis and the line segment passing through (x2, y min ) and perpendicular to the Y axis, the four line segments form a rectangle; the rectangle obtained by using the complete outline selection method for the real-time complete outline is marked as the real-time rectangle outline; this complete outline selection method can obtain the minimum rectangular outline; For specific details, please refer to Figure 6 As shown, get an xi The maximum coordinates of the value are: (6, 5), get an x i The minimum value is: (2, 5), get a y i The maximum coordinates of y are: (4, 16), get i The coordinates of the minimum value are: (4, 6); then obtain the line segment passing through (6, 5) and perpendicular to the X-axis, the line segment passing through (2, 5) and perpendicular to the X-axis, the line segment passing through (4, 16) and perpendicular to the Y-axis, and the line segment passing through (4, 6) and perpendicular to the Y-axis. The four line segments form a rectangle; Step S302, obtaining the length and width of the real-time rectangular outline, marked as Cs and Ks respectively; Step S303, calculate the real-time aspect ratio as: Qs=Cs / Ks; wherein Qs is the real-time aspect ratio; the principle of setting the real-time aspect ratio: due to the morphological characteristics of mesenchymal stem cells, that is, a long strip shape, the real-time aspect ratio is equal to data greater than 1, and other differences in cell morphology, such as hematopoietic stem cells are circular, that is, the real-time aspect ratio is close to 1, so the historical aspect ratio range of mesenchymal stem cells can be used to determine whether the cells are mesenchymal stem cells. In practice, take a data as an example: obtain the length and width of the real-time rectangular outline, marked as Cs=10 and Ks=4 respectively, and calculate the real-time aspect ratio: Qs=Cs / Ks=2.5.

[0020] Step S4, obtaining a range of historical aspect ratios based on the first number of historical mesenchymal stem cells; Step S4 includes the following sub-steps: Step S401, obtaining a first number of historical mesenchymal stem cells, and obtaining a complete historical profile of the historical mesenchymal stem cells; Step S402, for any historical complete contour, obtain a rectangle from the historical complete contour using a complete contour box selection method, and mark it as a historical rectangular contour; Step S403, obtaining the length and width of the historical rectangular outline, marked as Cl and Kl respectively; calculating the historical aspect ratio: Ql=Cl / Kl; wherein Ql is the historical aspect ratio; Step S404, see Figure 7 As shown, the historical aspect ratio is evenly divided into multiple intervals, marked as historical aspect ratio intervals, and the frequency of the historical aspect ratio of each aspect ratio interval is counted, marked as the interval historical aspect ratio frequency; with the historical aspect ratio as the X-axis and the interval historical aspect ratio frequency as the Y-axis, a histogram is drawn, marked as the historical aspect ratio histogram; a historical aspect ratio histogram is established, and the distribution range of the historical aspect ratio can be intuitively observed through the historical aspect ratio histogram; Step S405, counting the number of historical aspect ratio intervals, marked as Z1; sorting the interval historical aspect ratio frequencies from small to large, setting a sequence number corresponding to each interval historical aspect ratio frequency, and the sequence number is an integer starting from 1; Step S406, determine whether w1*Z1 is an integer. If so, set the interval historical length-width ratio frequency of the sequence number w1*Z1 as the first ratio number, marked as T1; if not, calculate the average of the interval historical length-width ratio frequencies corresponding to the sequence numbers on both sides of w1*Z1 as the first ratio number; w1 is the first ratio; determine whether w2*Z1 is an integer. If so, set the interval historical length-width ratio frequency of the sequence number w2*Z1 as the second quantile, marked as T2; if not, calculate the quantile The average value of the interval historical length-width ratio frequency corresponding to the serial numbers on both sides of w2*Z1 is the second quantile; w2 is the first ratio, w2=[(1 / w1)-1]*w1; w1 and w2 are both between 0 and 1, and the optimal setting w1=0.25 is: T1 is the interval historical length-width ratio frequency at one-fourth of the interval arranged from small to large, and T2 is the interval historical length-width ratio frequency at three-fourths of the interval arranged from small to large; obtain the distribution range of the interval historical length-width ratio frequency of T1 and T2 to obtain the small threshold; Step S407, set the too small threshold value to: t1=T1-r*(T2-T1); where t1 is the too small threshold value, and r is the range constant; where r can be set according to the frequency distribution of the historical length-width ratio of the interval. Here, for the part with a larger frequency of the historical length-width ratio of the interval to be retained, the r value can be set to be smaller; In practice, the number of historical length-to-width ratio intervals is counted as Z1=6, and the interval historical length-to-width ratio frequencies are sorted from small to large, namely 50,000, 60,000, 900,000, 920,000, 1.02 million and 1.47 million. The historical length-to-width ratio frequency of each interval is set to correspond to a sequence number: 1, 2, 3, 4, 5 and 6. It is determined that 0.25*6 is not an integer, and the interval historical length-to-width ratio frequencies corresponding to the sequence numbers on both sides of 1.5 are 60,000 and 900,000, namely T1= (6+90) / 2=480,000; calculate w2=[(1 / w1)-1]*w1=0.75; determine that 0.75*6 is not an integer, and obtain the interval historical length-width ratio frequency corresponding to the sequence numbers on both sides of 4.5, which is 920,000 and 1,020,000, that is, T1=(92+102) / 2=970,000, and set the too small threshold as: t1=T1-r*(T2-T1)=48-0.75*(97-48)=112,500; Step S408: Determine whether all the historical aspect ratio frequencies of the intervals are less than the过小threshold. If so, mark the historical aspect ratio intervals corresponding to the historical aspect ratio frequencies less than the过小threshold as过小intervals, delete the过小intervals from the historical aspect ratio histogram, and mark the historical aspect ratio histogram after deleting the过小intervals as the final aspect ratio histogram. If not, directly mark the historical aspect ratio histogram as the final aspect ratio histogram; Step S409: Obtain the minimum and maximum values of the historical aspect ratio intervals in the X-axis data of the final aspect ratio histogram, and mark them as Qmin and Qmax respectively; In a specific actual situation, it is obtained that the historical aspect ratio frequency of the interval is less than the过小threshold = 112,500. Mark the historical aspect ratio intervals corresponding to the frequencies of 50,000 and 60,000 as过小intervals, delete the过小intervals from the historical aspect ratio histogram, and mark the historical aspect ratio histogram after deleting the过小intervals as the final aspect ratio histogram.

[0021] Step S5: Obtain the qualified real-time aspect ratio based on the ranges of the real-time aspect ratio and the historical aspect ratio, and obtain the number of qualified mesenchymal stem cells based on the number of all qualified real-time aspect ratios; Step S5 includes the following sub-steps: Step S501: Determine whether all the real-time aspect ratios satisfy Qmin < Qs < Qmax. If so, retain the real-time rectangular contour. If not, delete the real-time rectangular contour; Obtain the number of all real-time rectangular contours, and mark it as the number of qualified mesenchymal stem cells; Count the number of qualified mesenchymal stem cells to facilitate corresponding processing and operations in a timely manner; In a specific actual situation, taking a data as an example: It can be seen from the final aspect ratio histogram that Qmin and Qmax are 1.2 and 2.8. Since Qmin = 1.2 < Qs = 2.5 < Qmax = 2.8, the real-time rectangular contour is retained.

[0022] Embodiment 2, a schematic diagram of the structure of an electronic device, the electronic device may include: a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other through the communication bus. The memory stores computer-readable instructions, the processor may call the instructions in the memory, and when the computer-readable instructions are executed by the processor, the steps in a stem cell culture image recognition detection method are executed to achieve the following functions: obtaining a culture image of mesenchymal stem cells based on a microscope and marking it as a stem cell image; graying the stem cell image to obtain a stem cell grayscale image, binarizing the stem cell grayscale image to obtain a stem cell binarization image, and obtaining a real-time complete outline based on the stem cell binarization image; obtaining a real-time rectangular outline of the real-time complete outline based on the complete outline selection method, and obtaining a real-time aspect ratio based on the real-time rectangular outline; obtaining a range of historical aspect ratios based on the first number of historical mesenchymal stem cells; obtaining a qualified real-time aspect ratio based on the real-time aspect ratio and the range of historical aspect ratios, and obtaining the number of qualified mesenchymal stem cells based on the number of all qualified real-time aspect ratios.

[0023] In addition, the logic instructions in the above-mentioned memory can be implemented in the form of software functional units and can be stored in a computer-readable storage medium when sold or used as an independent product. Based on this understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0024] Embodiment 3, the present application also provides a computer program product, the computer program product includes a computer program stored on a computer-readable storage medium, the computer program includes program instructions, when the program instructions are executed by the computer, the computer can execute a stem cell culture image recognition and detection method provided by the above methods, the method comprising: obtaining a culture image of mesenchymal stem cells based on a microscope, marking it as a stem cell image; graying the stem cell image to obtain a stem cell grayscale image, binarizing the stem cell grayscale image to obtain a stem cell binarization image, and obtaining a real-time complete outline based on the stem cell binarization image; obtaining a real-time rectangular outline of the real-time complete outline based on a complete outline frame selection method, and obtaining a real-time aspect ratio based on the real-time rectangular outline; obtaining a range of historical aspect ratios based on a first number of historical mesenchymal stem cells; obtaining a qualified real-time aspect ratio based on the real-time aspect ratio and the range of historical aspect ratios, and obtaining the number of qualified mesenchymal stem cells based on the number of all qualified real-time aspect ratios.

[0025] Embodiment 4, the present application also provides a computer-readable storage medium, the present application provides a storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps in the above stem cell culture image recognition and detection method are executed to achieve the following functions: obtaining a culture image of mesenchymal stem cells based on a microscope and marking it as a stem cell image; graying the stem cell image to obtain a stem cell grayscale image, binarizing the stem cell grayscale image to obtain a stem cell binarization image, and obtaining a real-time complete outline based on the stem cell binarization image; obtaining a real-time rectangular outline of the real-time complete outline based on a complete outline selection method, and obtaining a real-time aspect ratio based on the real-time rectangular outline; obtaining a range of historical aspect ratios based on a first number of historical mesenchymal stem cells; obtaining a qualified real-time aspect ratio based on the real-time aspect ratio and the range of historical aspect ratios, and obtaining the number of qualified mesenchymal stem cells based on the number of all qualified real-time aspect ratios.

[0026] Through the description of the above implementation methods, the embodiments of the present invention can be provided as methods, systems or computer program products. Based on such an understanding, the above technical solutions can be essentially or partly contributed to the prior art in the form of software products, which can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, an optical disk, etc., and include several instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0027] In the embodiments provided in the present application, it should be understood that the disclosed system or method can be implemented in other ways. The embodiments described above are merely schematic. For example, the division of modules or units is only a logical function division. There may be other division methods in actual implementation. For example, multiple modules or units can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some communication interfaces, and the indirect coupling or communication connection of systems, modules and units can be electrical, mechanical or other forms.

[0028] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A stem cell culture image recognition detection method, characterized in that: The method comprises the following steps: obtaining a culture image of mesenchymal stem cells based on a microscope and marking the image as a stem cell image; Gray-processing the stem cell image to obtain a stem cell grayscale image, binarizing the stem cell grayscale image to obtain a stem cell binary image, and obtaining a real-time complete contour based on the stem cell binary image; Based on the complete outline selection method, a real-time rectangular outline of the real-time complete outline is obtained, and based on the real-time rectangular outline, a real-time aspect ratio is obtained; Obtaining a range of historical aspect ratios based on the first number of historical mesenchymal stem cells; A qualified real-time aspect ratio is obtained based on the range of the real-time aspect ratio and the historical aspect ratio, and the number of qualified mesenchymal stem cells is obtained based on the number of all qualified real-time aspect ratios.

2. A stem cell culture image recognition detection method according to claim 1, characterized in that: Grayscale processing of the stem cell image to obtain a stem cell grayscale image includes the following sub-steps: Obtain the R, G and B values ​​of the RGB value of each pixel of the stem cell image; Use the grayscale conversion formula to convert the RGB value into a grayscale value. The grayscale conversion formula is: Hd=p1*R+p2*G+p3*B; Hd is the grayscale value of the pixel in the stem cell image, and p1, p2 and p3 are the R, G and B weights respectively.

3. A stem cell culture image recognition detection method according to claim 2, characterized in that: Binarization of the stem cell grayscale image to obtain a stem cell binary image includes the following sub-steps: Divide the grayscale value of 0-255 into f intervals evenly, marked as grayscale intervals; Count the frequency of the grayscale values ​​in each grayscale interval respectively, and mark it as grayscale interval frequency; Draw a histogram with grayscale value as X-axis and grayscale interval frequency as Y-axis, marked as grayscale histogram; Add up the frequencies of all grayscale intervals to get the total frequency, marked as J; Mark the intervals with grayscale interval frequency less than J / f as smaller areas, and mark the intervals other than the smaller areas as larger areas. Determine whether there are larger areas on both sides of the smaller area. If so, mark the smaller area as the final area. The midpoint value of the abscissa of the final area is set as the background threshold; the grayscale value of the pixel points greater than or equal to the background threshold is set to 255, and the grayscale value of the pixel points less than the background threshold is set to 0.

4. A stem cell culture image recognition detection method according to claim 3, characterized in that: Obtaining a real-time complete outline based on the stem cell binary image includes the following sub-steps: If a pixel with a gray value of 255 is adjacent to a pixel with a gray value of 0, the pixel with a gray value of 0 is marked as a stem cell edge pixel; All pixels except those at the edge of stem cells were set to 255; Obtain any pixel point at the edge of a stem cell, mark it as the initial center pixel point, establish a 3x3 pixel area with the initial center pixel point as the center, mark it as the reference area, and the nine pixels in the reference area include the initial center pixel point and eight neighboring pixels around the initial center pixel point; Determine whether the eight neighboring pixels around the initial center pixel point contain a stem cell edge pixel point. If so, mark the stem cell edge pixel point adjacent to the initial center pixel point as an adjacent edge pixel point. If not, remove the initial center pixel point and continue to select the next arbitrary stem cell edge pixel point as the initial center pixel point. A new reference area is established again by taking any adjacent edge pixel point as the center pixel point of the reference area, obtaining the newly appeared stem cell edge pixel point from the new reference area and marking it as a new edge pixel point, and the new edge pixel point is used as the center pixel point of the new reference area to establish a new reference area again, and so on, and new reference areas are continuously established until there are no new stem cell edge pixels in the new reference area; Determine whether the pixels in the last established reference area include the initial central pixel. If it does, obtain the central pixels of all the reference areas and mark them as completed contours. If it does not include the initial central pixel, obtain the central pixel of the last established reference area and mark it as the final central pixel. Establish a 5x5 pixel area with the final central pixel as the center and mark it as the final reference area. Determine whether the final reference area includes the initial central pixel. If it does, obtain the pixels on the straight path between the initial central pixel and the final central pixel and mark them as newly added contour pixels. Set the grayscale value of the newly added contour pixels to 0, obtain the central pixels of all the reference areas and the newly added contour pixels and mark them as completed contours. If it does not include the initial central pixel, obtain the central pixels of all the reference areas and mark them as real-time incomplete contours. Complete contours and completed contours are collectively referred to as real-time complete contours. Remove real-time incomplete contours.

5. A stem cell culture image recognition detection method according to claim 4, characterized in that: Obtaining a real-time rectangular outline of a real-time complete outline based on the complete outline selection method, and obtaining a real-time aspect ratio based on the real-time rectangular outline includes the following sub-steps: The complete outline selection method includes: Establishing a real-time plane rectangular coordinate system, placing each individual real-time complete profile in the first quadrant; Set the coordinate point of the real-time complete contour in the real-time rectangular coordinate system to (x i ,y i ), get an x i The maximum coordinates of max , y1), get an x i The minimum value of x is: min , y2), get a y i The maximum coordinates are: (x1, y max ), get y i The coordinates of the minimum value are: (x2, y min Then get (x max , y1) and perpendicular to the X-axis, through (x min , y2) and perpendicular to the X-axis, through (x1, y max ) and perpendicular to the Y axis and the line segment passing through (x2, y min ) and perpendicular to the Y-axis, the four line segments form a rectangle; Mark the rectangle obtained by the real-time complete outline using the complete outline frame selection method as the real-time rectangle outline; Get the length and width of the real-time rectangular outline, marked as Cs and Ks respectively; The real-time aspect ratio is calculated as: Qs=Cs / Ks; where Qs is the real-time aspect ratio.

6. A stem cell culture image recognition detection method according to claim 5, characterized in that: Acquiring a range of historical aspect ratios based on the first number of historical mesenchymal stem cells comprises the following sub-steps: Obtain the first quantity of historical mesenchymal stem cells, obtain the complete historical profile of historical mesenchymal stem cells; For any historical complete contour, the historical complete contour is obtained into a rectangle by using the complete contour box selection method, and is marked as the historical rectangular contour; Get the length and width of the historical rectangular outline, marked as Cl and Kl respectively; The historical aspect ratio is calculated as: Ql=Cl / Kl; where Ql is the historical aspect ratio.

7. A stem cell culture image recognition detection method according to claim 6, characterized in that: Acquiring a range of historical aspect ratios based on the first number of historical mesenchymal stem cells also includes the following sub-steps: Evenly divide the historical aspect ratio into multiple intervals, marked as historical aspect ratio intervals, and count the frequency of the historical aspect ratio in each aspect ratio interval, marked as interval historical aspect ratio frequency; With the historical aspect ratio as the X-axis and the interval historical aspect ratio frequency as the Y-axis, a histogram is drawn and marked as the historical aspect ratio histogram.

8. A stem cell culture image recognition detection method according to claim 7, characterized in that: Acquiring a range of historical aspect ratios based on the first number of historical mesenchymal stem cells also includes the following sub-steps: Count the number of historical aspect ratio intervals and label it as Z1; Sort the historical aspect ratio frequencies of the intervals from smallest to largest, and assign an ordinal number to each historical aspect ratio frequency of the interval. The ordinal number is an integer starting from 1; Judge whether w1 * Z1 is an integer. If it is an integer, set the historical aspect ratio frequency of the interval with the serial number w1 * Z1 as the first ratio number and label it as T1; if it is not an integer, calculate the average value of the historical aspect ratio frequencies of the intervals corresponding to the serial numbers on both sides of w1 * Z1 as the first ratio number; where w1 is the first ratio; Judge whether w2 * Z1 is an integer. If it is an integer, set the historical aspect ratio frequency of the interval with the serial number w2 * Z1 as the second quantile and label it as T2; if it is not an integer, calculate the average value of the historical aspect ratio frequencies of the intervals corresponding to the serial numbers on both sides of w2 * Z1 as the second quantile; where w2 is the first ratio, w2 = [(1 / w1) - 1] * w1; Set the过小 threshold as: t1 = T1 - r * (T2 - T1); where t1 is the过小 threshold and r is the range constant.

9. A stem cell culture image recognition detection method according to claim 8, characterized in that: Obtaining the range of historical aspect ratios based on the first quantity of historical mesenchymal stem cells further includes the following sub-steps: Judge whether all the historical aspect ratio frequencies of the intervals are less than the过小 threshold. If so, label the historical aspect ratio intervals corresponding to the historical aspect ratio frequencies less than the过小 threshold as过小 intervals, delete the过小 intervals in the historical aspect ratio histogram, and label the historical aspect ratio histogram after deleting the过小 intervals as the final aspect ratio histogram; If not, directly label the historical aspect ratio histogram as the final aspect ratio histogram; Obtain the minimum and maximum values of the historical aspect ratio intervals in the X-axis data of the final aspect ratio histogram, and label them as Qmin and Qmax respectively.

10. A stem cell culture image recognition detection method according to claim 9, characterized in that: Obtaining qualified real-time aspect ratios based on the range of real-time aspect ratios and historical aspect ratios, and obtaining the number of qualified mesenchymal stem cells based on the number of all qualified real-time aspect ratios includes the following sub-steps: Judge whether all the real-time aspect ratios satisfy Qmin < Qs < Qmax. If so, retain the real-time rectangular contour; if not, delete the real-time rectangular contour; obtain the number of all real-time rectangular contours and label it as the number of qualified mesenchymal stem cells.

Images