CN118134802A - Stem cell storage-oriented cell distribution intelligent detection method and system - Google Patents
Stem cell storage-oriented cell distribution intelligent detection method and system Download PDFInfo
- Publication number
- CN118134802A CN118134802A CN202410443111.3A CN202410443111A CN118134802A CN 118134802 A CN118134802 A CN 118134802A CN 202410443111 A CN202410443111 A CN 202410443111A CN 118134802 A CN118134802 A CN 118134802A
- Authority
- CN
- China
- Prior art keywords
- cell
- sliding
- window
- pixel
- stem cell
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/70—Denoising; Smoothing
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0012—Biomedical image inspection
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/13—Edge detection
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10056—Microscopic image
- G06T2207/10061—Microscopic image from scanning electron microscope
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Radiology & Medical Imaging (AREA)
- Quality & Reliability (AREA)
- Image Analysis (AREA)
Abstract
The invention relates to the technical field of image enhancement, in particular to an intelligent cell distribution detection method and system for stem cell storage, which obtain an initial denoising window based on the image characteristics of an obtained stem cell gray level image so as to obtain a local cell area sliding each time; analyzing the cell detail characteristic distribution condition and the cell edge distribution rule degree of the local cell area, and determining a first window variation parameter and a second window variation parameter; the local cell area sliding each time is adaptively adjusted through the first window variable parameter and the second window variable parameter to obtain each final denoising area, and further a denoised stem cell image is obtained; and obtaining the distribution condition of the stem cells to be detected according to the denoised stem cell image. According to the invention, through self-adapting the size of the denoising area of the stem cell image, the defect of poor denoising effect of a global image-based denoising algorithm is overcome, and the accuracy of the intelligent detection result of the cell distribution stored by the stem cells is improved.
Description
Technical Field
The invention relates to the technical field of image enhancement, in particular to an intelligent cell distribution detection method and system for stem cell storage.
Background
The storage of stem cells is critical for medical research and clinical treatment, and the distribution of cells during the storage of stem cells is directly related to the storage quality and the effect of subsequent applications. The traditional manual detection of the cell distribution condition is low in efficiency, easy to make mistakes and difficult to meet the actual requirements; when the intelligent detection is carried out on the cell distribution condition, due to the influence of factors such as noise, illumination conditions, image distortion and the like of the image acquisition equipment, spots, stripes, disordered backgrounds or inconsistent pixel intensities can appear in the stem cell image, the quality of the obtained stem cell image is low, the cell boundary of the stem cell image with low quality is unclear, and the segmentation and positioning of cells are easily influenced. Therefore, the stem cell image needs to be subjected to denoising processing.
The stem cell image has a complex structure and more detail characteristics, the existing global image-based denoising algorithm is utilized to carry out image enhancement processing on the stem cell image, so that the detail blurring or detail loss of the processed stem cell image can be caused, namely the denoising effect on the stem cell image is poor, and the accuracy of the intelligent detection result of the cell distribution stored by the stem cell is further caused to be low.
Disclosure of Invention
In order to solve the technical problem that the accuracy of the cell distribution intelligent detection result of stem cell storage is low due to poor denoising effect of the existing global image-based denoising algorithm, the invention aims to provide a cell distribution intelligent detection method and system for stem cell storage, and the adopted technical scheme is as follows:
an embodiment of the invention provides an intelligent detection method for cell distribution of stem cell storage, which comprises the following steps:
acquiring a stem cell gray level image of stem cells to be detected;
Obtaining an initial denoising window of the stem cells to be detected according to the pixel value of each pixel point in the stem cell gray level image and the image size of the stem cell gray level image; sliding an initial denoising window on the stem cell gray level image according to a preset step length to obtain a local cell area sliding each time;
Analyzing the distribution condition of cell detail characteristics according to the pixel value of each pixel point in the local cell area of each sliding, and determining a first window variation parameter of each sliding;
Analyzing the distribution rule degree of the cell edges according to the local cell area of each sliding, and determining a second window variation parameter of each sliding;
according to the first window variable parameter and the second window variable parameter of each sliding, self-adaptively adjusting the local cell area of each sliding to obtain each final denoising area corresponding to the stem cell gray level image;
Denoising operation is carried out on each final denoising area, and a denoised stem cell image is obtained; and (3) analyzing the cell distribution according to the denoised stem cell image to obtain the distribution condition of the stem cells to be detected.
Further, the obtaining an initial denoising window of the stem cell to be detected according to the pixel value of each pixel point in the stem cell gray level image and the image size of the stem cell gray level image includes:
Calculating the pixel average value of all pixel points in the stem cell gray level image, taking the pixel average value as a judgment threshold value of the cell pixel points, and further selecting the pixel points with the pixel values larger than the judgment threshold value in the stem cell gray level image as the cell pixel points; taking other pixel points except the cell pixel points in the stem cell gray level image as background pixel points;
determining a first window size according to the ratio of all cell pixel points in the stem cell gray level image and the image size of the stem cell gray level image;
Parity judgment is carried out on the first window size, if the first window size is odd, the window formed by the first window size is directly used as an initial denoising window; if the first window size is even, odd processing is carried out on the first window size, and a window formed by the numerical values after the odd processing is used as an initial denoising window.
Further, the calculation formula of the first window size is:
Where w 1 is the first window size,/> N cell is the number of cell pixels in the stem cell gray scale image, N bg is the number of background pixels in the stem cell gray scale image, min is the minimum function, len is the length of the stem cell gray scale image, and wei is the width of the stem cell gray scale image.
Further, the analyzing the distribution of the cell detail features according to the pixel value of each pixel point in the local cell area of each sliding, and determining the first window variation parameter of each sliding includes:
for any one sliding, determining the number of all cell pixel points in the local cell area of the sliding, and further determining the pixel average value and the pixel variance corresponding to all the pixel points in the local cell area of the sliding;
and determining a first window variation parameter of the sliding according to the number of all cell pixels in the local cell region of the sliding, the number of all pixel points in the local cell region, the pixel average value and the pixel variance corresponding to all pixel points and the pixel value of each pixel point in the local cell region of the sliding.
Further, determining the first window variation parameter of the sliding according to the number of all the cellular pixels in the local cellular region, the pixel average value and the pixel variance corresponding to all the cellular pixels, and the pixel value of each cellular pixel in the local cellular region, where the first window variation parameter includes:
Wherein, C is a first window variable parameter of the sliding, n cell is the number of all the cell pixels in the local cell region of the sliding, m 2 is the number of all the pixel pixels in the local cell region of the sliding, m is the size of the local cell region of the sliding, i is the serial number of each pixel in the local cell region of the sliding, x i is the pixel value of the ith pixel in the local cell region of the sliding, μ is the average value of the pixels corresponding to all the pixel in the local cell region of the sliding, σ 2 is the variance corresponding to all the pixel in the local cell region of the sliding, i is an absolute function, exp is an exponential function based on a natural constant.
Further, the determining the second window variation parameter of each sliding according to the local cell area analysis cell edge distribution regularity of each sliding includes:
Performing edge detection on the local cell area sliding each time to obtain each edge in the local cell area sliding each time;
For any sliding, determining the curvature integral result of each edge in the local cell area of the sliding, and taking the average value of the curvature integral results of all edges as a cell edge distribution rule index of the local cell area of the sliding;
Calculating the average edge length of the local cell area and the average curvature of all edges when the sliding is performed last time, and taking the product of the average edge length and the average curvature of all edges as a standard rule index;
And determining the difference value between the cell edge distribution rule index and the reference rule index of the local cell area of the sliding, and normalizing the difference value to obtain a normalized numerical value serving as a second window variation parameter of the sliding.
Further, according to the first window variation parameter and the second window variation parameter of each sliding, adaptively adjusting the local cell area of each sliding to obtain each final denoising area corresponding to the stem cell gray level image, including:
For any one sliding, determining a window size adjustment value of the sliding according to a first window variable parameter and a second window variable parameter of the sliding;
Calculating the product of the window size adjustment value sliding for the time and the size of the initial denoising window, and adding the product and the size of the initial denoising window;
the added values are used as the target size of the sliding, and the area formed by the target size is used as the final denoising area.
Further, the determining the window size adjustment value of the sliding according to the first window variable parameter and the second window variable parameter of the sliding includes:
The product of the first window variable and the second window variable of the sliding is calculated, and then the value obtained by adding the product of the two window variable and the first window variable is used as the window size adjustment value of the sliding.
Further, the analysis of cell distribution according to the denoised stem cell image to obtain the distribution condition of the stem cells to be detected includes:
Performing edge detection on the denoised stem cell image to obtain the edges of each cell; connecting and repairing the edges of each cell to obtain the boundaries of each stem cell; and identifying and positioning the boundaries of each stem cell to obtain the distribution condition of the stem cells to be detected.
The embodiment of the invention also provides a cell distribution intelligent detection system for stem cell storage, which comprises a processor and a memory, wherein the processor is used for processing instructions stored in the memory so as to realize a cell distribution intelligent detection method for stem cell storage.
The invention has the following beneficial effects:
The invention provides a cell distribution intelligent detection method and system for stem cell storage, relates to the technical field of image enhancement, and is particularly applied to the field of cell distribution detection. Combining the pixel characteristics and the image size of the stem cell gray level image, the obtained initial denoising window is more in accordance with the integral cell distribution characteristics of the stem cell gray level image, the robustness of the initial denoising window is stronger, and meanwhile, reference data is provided for subsequent self-adaptive adjustment denoising areas; when the window variation parameters are quantized, the first window variation parameters and the second window variation parameters are determined from two angles of cell detail characteristic distribution conditions and cell edge distribution regularity, so that the cell detail characteristic conditions of the local cell area sliding each time can be comprehensively and completely known, and the reliability of each final denoising area obtained based on the first window variation parameters and the second window variation parameters can be improved.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions and advantages of the prior art, the following description will briefly explain the drawings used in the embodiments or the description of the prior art, and it is obvious that the drawings in the following description are only some embodiments of the invention, and other drawings can be obtained according to the drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flow chart of a method for intelligently detecting cell distribution for stem cell storage according to an embodiment of the invention;
FIG. 2 is a gray scale image of stem cells in an embodiment of the invention.
Detailed Description
In order to further describe the technical means and effects adopted by the present invention to achieve the preset purpose, the following detailed description is given below of the specific implementation, structure, features and effects of the technical solution according to the present invention with reference to the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" means that the embodiments are not necessarily the same. Furthermore, the particular features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
The specific conditions aimed by the invention can be as follows:
When the prior art is used for intelligently detecting stem cell images, spots, stripes, disordered backgrounds or inconsistent pixel intensities appear in the stem cell images due to the influence of factors such as noise, illumination conditions, image distortion and the like of image acquisition equipment, so that the definition of the boundaries of the stem cells is low, and the segmentation and positioning of cells are influenced. Therefore, the denoising processing needs to be performed on the stem cell image, and the denoising algorithm based on the global image may cause the details of the stem cells to be fuzzy and lost, so that the distribution condition of the stem cells cannot be accurately detected.
In order to overcome the defects of a denoising algorithm based on a global image and improve the accuracy of a cell distribution intelligent detection result, the embodiment of the invention provides a stem cell storage-oriented cell distribution intelligent detection method, which comprises the following steps as shown in fig. 1:
s1, acquiring a stem cell gray level image of a stem cell to be detected.
In this embodiment, the main purpose is to perform relevant cell distribution detection by using stem cell images that need to be stored, so as to implement intelligent monitoring and management of stem cell storage processes. Before intelligent detection of cell distribution, obtaining a stem cell image which needs to be subjected to intelligent detection of the distribution, namely, a stem cell image of stem cells to be detected. In acquiring stem cell images, this is achieved by a photographable microscope, wherein the microscope is capable of providing high resolution stem cell images to capture microstructure and detail features of stem cells; the stem cells in the stem cell gray scale image are not single but are present in plurality.
After obtaining the stem cell image of the stem cell to be detected, in order to facilitate the subsequent analysis of the stem cell image, the stem cell image needs to be subjected to gray-scale treatment, and a stem cell gray-scale image can be obtained, wherein the stem cell gray-scale image is shown in fig. 2.
The implementation method of the graying treatment comprises the following steps: the implementation process of the graying process is a prior art, and will not be described in detail here.
Thus, the embodiment obtains a stem cell gray scale image of stem cells to be detected.
S2, obtaining an initial denoising window of the stem cells to be detected according to the pixel value of each pixel point in the stem cell gray level image and the image size of the stem cell gray level image, and further obtaining a local cell area sliding each time.
The stem cell image is different from other types of images, has more details, and has the characteristics of irregular distribution and large-area stem cell group, and of course, there are cases where a large-area culture solution is free of stem cells. Therefore, regarding the image features of the stem cell image, if the stem cell image is scanned in a conventional manner, the possibility of losing details or amplifying noise as cell details easily occurs, so that the results of extracting and recognizing the distribution of the stem cell boundaries are affected. Therefore, the image information of the stem cells is utilized to self-adapt to the size of the local area of each denoising, and further, the stem cell image with the denoising operation is obtained.
The first step, obtaining an initial denoising window of the stem cells to be detected according to the pixel value of each pixel point in the stem cell gray level image and the image size of the stem cell gray level image.
In this embodiment, the size of the initial denoising window cannot be too large nor too small. The noise of the stem cell gray level image is relatively smaller, the initial denoising window cannot be too large, the ratio condition of the whole cells can represent the distribution condition of effective pixel points in the whole stem cell gray level image to a certain extent, the distribution condition of the effective pixel points can be taken as a consideration factor for defining the size of the initial denoising window, and the method has a certain inhibition effect on the quantification of the size of the initial denoising window to a certain extent; the smaller data in the length and width of the stem cell gray level image is selected as the reference of the size of the initial denoising window, so that the condition that the adaptability of the initial denoising window is low due to the influence of the image size can be prevented. The effective pixel points are cell pixel points.
The first substep calculates the average value of all pixels in the stem cell gray level image as the judgment threshold value of the cell pixels, and then selects the pixels with the pixel value larger than the judgment threshold value in the stem cell gray level image as the cell pixels, and uses the other pixels except the cell pixels in the stem cell gray level image as the background pixels. In this embodiment, the gray value of the cell pixel is generally higher than the gray value of the background pixel, so that all the cell pixels in the stem cell gray image can be selected by the determination threshold of the cell pixel obtained from the stem cell gray image. It should be noted that, the cell pixel points are selected to measure the density of the cell pixel points in the whole stem cell gray level image, and the more densely the cell pixel points are distributed, the smaller the initial denoising window should be in order to preserve the detail characteristics of the stem cells. And a second substep, determining the first window size according to the ratio of all cell pixels in the stem cell gray level image and the image size of the stem cell gray level image. As an example, the calculation formula of the first window size may be: Where w 1 is the first window size,/> For the downward rounding function Ncell is the number of cell pixels in the stem cell gray image, nbg is the number of background pixels in the stem cell gray image, min is the minimum function, len is the length of the stem cell gray image, and wei is the width of the stem cell gray image. In the calculation formula of the first window size, the first window size can represent the size of an initial denoising window, and the larger the first window size is, the larger the initial denoising window is; /(I)Can characterize the duty ratio of all cell pixel points in the stem cell gray level image,/>The smaller the size, the better the ratio condition of all cell pixel points in the stem cell gray level image is, the smaller the first window size is; min (len, wei) can characterize the baseline of the initial denoising window size, the larger the min (len, wei) the larger the first window size should be; pair/>Performing a fourth-order opening process to reduce the size of the first window to a certain extent; pair/>The rounding process is performed to avoid that the initial denoising window size cannot be reasonably determined when the value is a fraction, because: the size of the stem cell gray level image is composed of a plurality of pixel points, and if the size of the first window size is a decimal, an initial denoising window cannot be constructed. And a third sub-step, performing parity judgment on the first window size to obtain an initial denoising window of the stem cells to be detected. In this embodiment, the side length of the conventional sliding window is an odd number, so after the first window size is obtained, parity determination may be performed on the first window size to satisfy the reference condition of the sliding window size, and the specific implementation process may include: if the first window size is odd, directly taking a window formed by the first window size as an initial denoising window; if the first window size is even, odd processing is performed on the first window size, a window formed by the values after the odd processing is used as an initial denoising window, and the size of the initial denoising window can be w 2×w2. Wherein the sides of the initial denoising window are equal. As an example, the expression for parity determination for the first window size may be: Where w 2 is the first window size after parity determination, w 1 is the first window size, mod is a remainder function, w 1 mod 2=0 represents the first window size as even number, and w 1 mod 2=1 represents the first window size as odd number.
And secondly, sliding an initial denoising window on the stem cell gray level image according to a preset step length to obtain a local cell area sliding each time.
In this embodiment, on the stem cell gray level image, the initial denoising window may be slid in the sliding sequence from left to right and from top to bottom until the whole stem cell gray level image is traversed, and the area corresponding to the initial denoising window after each sliding is taken as the local cell area, so as to obtain the local cell area for each sliding. The preset step length of each sliding can be set to 1 pixel point, and an implementer can set the preset step length of each sliding according to specific practical conditions without specific limitation.
Thus far, the present embodiment obtains a local cell area for each sliding of stem cells to be detected. The obtained initial denoising window is more in accordance with the integral cell distribution characteristics of the stem cell gray level image by combining the pixel characteristics and the image size of the stem cell gray level image, the robustness of the initial denoising window is stronger, and meanwhile, reference data is provided for subsequent self-adaptive adjustment denoising areas.
S3, analyzing the distribution condition of cell detail characteristics according to the pixel value of each pixel point in the local cell area of each sliding, and determining a first window variation parameter of each sliding.
First, it should be noted that the first window variable is an index obtained by analyzing the distribution of cell detail characteristics, and the first window variable may be used to adaptively adjust the size of the local cell area. The detail characteristics of the local cell area are distributed more, and the window needs to be reduced to ensure the accuracy of image denoising; the detail characteristic distribution of the local cell area is less, and a window needs to be enlarged to ensure the image denoising efficiency.
In this embodiment, the determining process of the first window variation parameter of each sliding is consistent, and in order to reduce unnecessary description, taking any sliding as an example to determine the first window variation parameter of the sliding, a specific implementation process may include:
The first step is to determine the number of all the cell pixels in the local cell region of the sliding, and then determine the pixel average value and the pixel variance corresponding to all the pixel pixels in the local cell region of the sliding.
In this embodiment, the number of pixels with a pixel value greater than the determination threshold of the cell pixels in the local cell region of the sliding is counted to be the number of all the cell pixels in the local cell region of the sliding; the process of calculating the pixel mean and pixel variance is prior art and will not be described in detail herein. Wherein, the pixel value refers to the gray value of the pixel point.
It should be noted that, the number of all the cellular pixels, the pixel average value and the pixel variance corresponding to all the cellular pixels in the local cellular region of the sliding are determined to provide data support for the subsequent determination of the first window variation parameter.
And a second step of determining a first window variation parameter of the sliding according to the number of all the cell pixels in the local cell region of the sliding, the number of all the pixel pixels in the local cell region, the pixel average value and the pixel variance corresponding to all the pixel pixels and the pixel value of each pixel in the local cell region of the sliding.
In this embodiment, the cell characteristics of the local cell area formed by the initial denoising window during the sliding process are different, so that there is a possibility that the size of the local cell area is increased or decreased, and the first window variation parameter of the sliding is determined by fusing a plurality of different pixel characteristics.
As an example, the calculation formula of the first window variation parameter of the sliding may be:
Wherein, C is a first window variable parameter of the sliding, n cell is the number of all the cell pixels in the local cell region of the sliding, m 2 is the number of all the pixel pixels in the local cell region of the sliding, m is the size of the local cell region of the sliding, i is the serial number of each pixel in the local cell region of the sliding, x i is the pixel value of the ith pixel in the local cell region of the sliding, μ is the average value of the pixels corresponding to all the pixel in the local cell region of the sliding, σ 2 is the variance corresponding to all the pixel in the local cell region of the sliding, i is an absolute function, exp is an exponential function based on a natural constant.
In the calculation formula of the first window variation parameter,The method can be used for controlling the enlargement and the reduction of the local cell area of the sliding, and when the number of the cell pixel points in the local cell area is more than half, the cell detail characteristics of the local cell area of the sliding are more, and the local cell area needs to be reduced to ensure the denoising accuracy; similarly, when the background pixel points in the local cell area are more, the detail characteristics of the cells in the local cell area which slide for the time are less, and the local cell area needs to be enlarged to improve the denoising efficiency; /(I)The distribution of pixel values in the local cell region of the slip can be characterized,/>The smaller the pixel value distribution is, the more uniform the pixel value in the local cell area of the sliding is, the more detail features are, and the more the local cell area of the sliding is required to be subjected to shrinking treatment; exp (-) can be used for realizing the negative correlation processing of the normalization of the data, and an implementer can set a normalization function according to specific practical conditions, and the value range of the first window variation parameter is between 0 and 1.
Thus far, the present embodiment obtains the first window variation parameter for each sliding.
S4, determining a second window variation parameter of each sliding according to the distribution rule degree of the cell edge of the local cell area analysis cell of each sliding.
It should be noted that the second window variation parameter is an index obtained by analyzing an obvious distribution rule existing at the cell edge, and the second window variation parameter may be used to adjust the first window variation parameter to obtain a more accurate window size adjustment value. Because some detection errors may occur in the edge detection, interference edges of non-cell edges are generated in the stem cell gray level image, and the interference edges affect the extraction of cell characteristics, so that errors exist in the self-adaptive adjustment of local cell areas.
In this embodiment, the calculation process of the second window variation parameter of each sliding is consistent, and in order to reduce unnecessary description, taking any sliding as an example, determining the second window variation parameter of the sliding may include:
In the first step, edge detection is performed on the local cell area of each slide, and each edge in the local cell area of each slide is obtained.
In this embodiment, the edge detection technique is used to detect the edges of the local cell area in each slide, so that each edge in the local cell area in each slide may be obtained, where a cell edge and an interference edge may exist.
Edge detection techniques include, but are not limited to: the implementation process of edge detection is the prior art, and is not in the scope of the present invention, and detailed description is not provided herein.
And secondly, determining the curvature integral result of each edge in the sliding local cell area, and taking the average value of the curvature integral results of all edges as a cell edge distribution rule index of the sliding local cell area.
In this embodiment, the cell edge distribution rule index may be used to measure the degree of cell edge distribution rule of the local cell region sliding for this time, and the greater the cell edge distribution rule index, the stronger the edge distribution rule of the local cell region, which indicates that the more obvious the cell detail features are. The existence of cells in the local cell area can be estimated through the curvature integral result of the edge, and data support is provided for the follow-up determination of the second window variation parameter.
The curvature may represent the bending degree of the edge, and the curvature of each edge pixel point on the edge may be used as the weight of the corresponding edge pixel point to extract the whole edge feature in an integral manner, and the calculation process of the curvature is the prior art and will not be described in detail here.
As an example, the calculation formula of the cell edge distribution rule index of the local cell region of the sliding may be:
Wherein c is a cell edge distribution rule index of the local cell region of the sliding, G is the number of edges in the local cell region of the sliding, G is the number of edges in the local cell region of the sliding, l g is the number of edge pixel points on the G-th edge in the local cell region of the sliding, k g (x) is a curvature function of the G-th edge in the local cell region of the sliding,/> The curvature of the g-th edge in the local cell region for this sliding is integrated.
And thirdly, calculating the average edge length of the local cell area and the average curvature of all edges when the sliding is performed last time, and taking the product of the average edge length and the average curvature of all edges as a standard rule index.
In this embodiment, the average value of all edge lengths in the local cell area at the last sliding of the sliding is determined first, and is recorded as the average edge length, and the edge length refers to the number of edge pixel points on the edge; the average value of all edge curvatures in the local cell region at the last sliding of the sliding was determined again and recorded as the average curvature.
It should be noted that, the reference rule index may represent the edge distribution rule condition of the local cell area during the last sliding of the sliding, and the edge distribution rule condition of the local cell area during the last sliding of the sliding is taken as comparison data, which may be used to measure the difference between the edge distribution rules of two adjacent sliding, where the reference rule index is a parameter in the process of subsequently determining the second window variation parameter.
And step four, determining the difference between the cell edge distribution rule index and the reference rule index of the local cell area of the sliding, and normalizing the difference to obtain a normalized value which is used as a second window variation parameter of the sliding.
In this embodiment, the larger the difference between the cell edge distribution rule index and the reference rule index, the higher the rule degree of the cell edge distribution in the local cell area of the sliding is, the more the first window variation parameter needs to be adjusted, that is, the greater the degree of enlarging or reducing the local cell area is. The normalization process is to ensure that the value of the second window variable ranges from 0 to 1.
The implementation method of normalization processing includes, but is not limited to, linear normalization, mean normalization, nonlinear normalization, decimal scaling normalization, etc., and the implementation process of normalization processing is the prior art and is not described in detail herein.
Thus far, the present embodiment obtains the second window variation parameter for each sliding.
When the window variation parameters are quantized, the first window variation parameters and the second window variation parameters are determined from two angles of cell detail characteristic distribution conditions and cell edge distribution regularity, so that the cell detail characteristic conditions of the local cell area sliding each time can be comprehensively and completely known, and the reliability of each final denoising area obtained based on the first window variation parameters and the second window variation parameters can be improved.
S5, according to the first window variable parameter and the second window variable parameter of each sliding, self-adaptively adjusting the local cell area of each sliding to obtain each final denoising area corresponding to the stem cell gray level image.
After obtaining the window variation parameters for adjusting the size of the local cell region, adaptive adjustment is required to be performed on the local cell region sliding each time to obtain each final denoising region where denoising is required.
It should be noted that, for the first local cell area corresponding to the stem cell gray level image, that is, the first local cell area located at the upper left corner of the stem cell gray level image, there is no local cell area sliding last before the local cell area, and only the first window variation parameter of the local cell area needs to be analyzed when the local cell area is analyzed.
In this embodiment, the determining process of each final denoising region is consistent, and in order to reduce unnecessary descriptions, taking determining any one final denoising region as an example, the specific implementation process may include:
and a first step of determining a window size adjustment value of the sliding according to the first window variable parameter and the second window variable parameter of the sliding.
The first window variable parameter and the second window variable parameter are fused together to obtain a final window size adjustment value, wherein the window size adjustment value is obtained by adjusting the first window variable parameter by utilizing the second window variable parameter on the basis of the first window variable parameter, and the sliding local cell area has the corresponding window size adjustment value.
Specifically, the product of the first window variable and the second window variable of the sliding is calculated, and then the value obtained by adding the product of the two window variable and the first window variable is used as the window size adjustment value of the sliding.
And secondly, calculating the product of the window size adjustment value sliding for the time and the size of the initial denoising window, and adding the product and the size of the initial denoising window. The added values are used as the target size of the sliding, and the area formed by the target size is used as the final denoising area.
The target size is based on the first window size, and the target size is floated up and down to achieve adaptive variation of the first window size, so that a final denoising region is obtained, and the side lengths of the final denoising region are equal.
Thus, the present embodiment obtains each final denoising region corresponding to the stem cell gray scale image.
S6, denoising operation is carried out on each final denoising area, and a denoised stem cell image is obtained; and (3) analyzing the cell distribution according to the denoised stem cell image to obtain the distribution condition of the stem cells to be detected.
And in the first step, denoising operation is carried out on each final denoising area, and a denoised stem cell image is obtained.
In this embodiment, the denoising filter operation is performed on each final denoising region by using gaussian filtering, so as to obtain a denoised stem cell image. The implementation process of the gaussian filtering is the prior art, and is not in the scope of the present invention, and will not be described in detail here.
And secondly, analyzing cell distribution according to the denoised stem cell image to obtain the distribution condition of the stem cells to be detected.
It should be noted that, after the denoising operation on the stem cell gray level image is completed, in order to analyze the cell distribution stored in the stem cell, the edge detection technology and the cell identification technology are utilized to identify and detect the cells in the denoised stem cell image, so as to facilitate the subsequent analysis and calculation of the cell distribution, and the specific implementation process may include:
Firstly, performing edge detection on the denoised stem cell image by using a Canny edge detection algorithm to obtain each cell edge. The implementation process of the Canny edge detection algorithm is the prior art, and is not in the scope of the present invention, and will not be described in detail here.
Secondly, after edge detection, broken or discontinuous edges may occur, and in order to higher identify cell boundaries, each cell edge is connected and repaired to obtain each complete stem cell boundary.
It should be noted that the connection and repair of cell edges is a common image processing technique in image processing and computer vision, such as edge filling, texture-based repair, and image inpainting. The implementation process of connecting and repairing the cell edges is the prior art, and is not in the scope of the present invention, and will not be described in detail here.
Finally, after obtaining the boundary of each complete stem cell, carrying out positioning analysis on each stem cell by utilizing boundary position information, determining the distribution condition of the stem cells to be detected, and marking and recording the distribution condition of the stem cells to be detected, thereby obtaining the distribution condition of the stem cells to be detected.
It should be noted that, the cell distribution may be obtained by using a statistical method according to basic information of each cell boundary, and the specific implementation process is the prior art and will not be described in detail herein.
Thus, the embodiment completes the intelligent detection of the cell distribution of the stem cells.
The embodiment of the invention also provides a cell distribution intelligent detection system for stem cell storage, which comprises a processor and a memory, wherein the processor is used for processing instructions stored in the memory so as to realize a cell distribution intelligent detection method for stem cell storage.
In summary, according to the cell distribution intelligent detection method and system for stem cell storage, when the stem cell image is subjected to denoising, the existing global denoising algorithm is not directly utilized to perform corresponding denoising operation, but local information of the stem cell image is utilized to perform self-adaptive adjustment on a local area, so that the time required by denoising of the stem cell image is saved to a certain extent while the cell detail is reserved. The invention is beneficial to enhancing the denoising effect of the denoising algorithm on the stem cell image, and further improves the accuracy of the intelligent detection result of the cell distribution stored by the stem cells.
The above embodiments are only for illustrating the technical solution of the present invention, and not for limiting the same; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit of the invention and are intended to be included within the scope of the invention.
Claims (10)
1. The intelligent cell distribution detection method for stem cell storage is characterized by comprising the following steps of:
acquiring a stem cell gray level image of stem cells to be detected;
Obtaining an initial denoising window of the stem cells to be detected according to the pixel value of each pixel point in the stem cell gray level image and the image size of the stem cell gray level image; sliding an initial denoising window on the stem cell gray level image according to a preset step length to obtain a local cell area sliding each time;
Analyzing the distribution condition of cell detail characteristics according to the pixel value of each pixel point in the local cell area of each sliding, and determining a first window variation parameter of each sliding;
Analyzing the distribution rule degree of the cell edges according to the local cell area of each sliding, and determining a second window variation parameter of each sliding;
according to the first window variable parameter and the second window variable parameter of each sliding, self-adaptively adjusting the local cell area of each sliding to obtain each final denoising area corresponding to the stem cell gray level image;
Denoising operation is carried out on each final denoising area, and a denoised stem cell image is obtained; and (3) analyzing the cell distribution according to the denoised stem cell image to obtain the distribution condition of the stem cells to be detected.
2. The intelligent detection method for cell distribution of stem cell storage according to claim 1, wherein the obtaining an initial denoising window of stem cells to be detected according to the pixel value of each pixel point in the stem cell gray level image and the image size of the stem cell gray level image comprises:
Calculating the pixel average value of all pixel points in the stem cell gray level image, taking the pixel average value as a judgment threshold value of the cell pixel points, and further selecting the pixel points with the pixel values larger than the judgment threshold value in the stem cell gray level image as the cell pixel points; taking other pixel points except the cell pixel points in the stem cell gray level image as background pixel points;
determining a first window size according to the ratio of all cell pixel points in the stem cell gray level image and the image size of the stem cell gray level image;
Parity judgment is carried out on the first window size, if the first window size is odd, the window formed by the first window size is directly used as an initial denoising window; if the first window size is even, odd processing is carried out on the first window size, and a window formed by the numerical values after the odd processing is used as an initial denoising window.
3. The intelligent detection method for cell distribution of stem cell storage according to claim 2, wherein the calculation formula of the first window size is:
Where w 1 is the first window size,/> N cell is the number of cell pixels in the stem cell gray scale image, N bg is the number of background pixels in the stem cell gray scale image, min is the minimum function, len is the length of the stem cell gray scale image, and wei is the width of the stem cell gray scale image.
4. The intelligent detection method for cell distribution of stem cell-oriented storage according to claim 2, wherein the analyzing cell detail characteristic distribution according to the pixel value of each pixel point in the local cell area of each sliding, determining the first window variation parameter of each sliding, comprises:
for any one sliding, determining the number of all cell pixel points in the local cell area of the sliding, and further determining the pixel average value and the pixel variance corresponding to all the pixel points in the local cell area of the sliding;
and determining a first window variation parameter of the sliding according to the number of all cell pixels in the local cell region of the sliding, the number of all pixel points in the local cell region, the pixel average value and the pixel variance corresponding to all pixel points and the pixel value of each pixel point in the local cell region of the sliding.
5. The method for intelligently detecting cell distribution for stem cell storage according to claim 4, wherein determining the first window variation parameter of the sub-sliding according to the number of all cell pixels in the local cell region of the sub-sliding, the number of all pixel pixels in the local cell region, the pixel average value and the pixel variance corresponding to all pixel pixels, and the pixel value of each pixel in the local cell region of the sub-sliding comprises:
Wherein, C is a first window variable parameter of the sliding, n cell is the number of all the cell pixels in the local cell region of the sliding, m 2 is the number of all the pixel pixels in the local cell region of the sliding, m is the size of the local cell region of the sliding, i is the serial number of each pixel in the local cell region of the sliding, x i is the pixel value of the ith pixel in the local cell region of the sliding, μ is the average value of the pixels corresponding to all the pixel in the local cell region of the sliding, σ 2 is the variance corresponding to all the pixel in the local cell region of the sliding, i is an absolute function, exp is an exponential function based on a natural constant.
6. The intelligent detection method for cell distribution of stem cell-oriented storage according to claim 1, wherein the determining the second window variation parameter of each sliding according to the degree of cell edge distribution regularity of the local cell area analysis of each sliding comprises:
Performing edge detection on the local cell area sliding each time to obtain each edge in the local cell area sliding each time;
For any sliding, determining the curvature integral result of each edge in the local cell area of the sliding, and taking the average value of the curvature integral results of all edges as a cell edge distribution rule index of the local cell area of the sliding;
Calculating the average edge length of the local cell area and the average curvature of all edges when the sliding is performed last time, and taking the product of the average edge length and the average curvature of all edges as a standard rule index;
And determining the difference value between the cell edge distribution rule index and the reference rule index of the local cell area of the sliding, and normalizing the difference value to obtain a normalized numerical value serving as a second window variation parameter of the sliding.
7. The intelligent detection method for cell distribution of stem cell storage according to claim 1, wherein the step of adaptively adjusting the local cell area of each sliding according to the first window variable parameter and the second window variable parameter of each sliding to obtain each final denoising area corresponding to the stem cell gray level image comprises the following steps:
For any one sliding, determining a window size adjustment value of the sliding according to a first window variable parameter and a second window variable parameter of the sliding;
Calculating the product of the window size adjustment value sliding for the time and the size of the initial denoising window, and adding the product and the size of the initial denoising window;
the added values are used as the target size of the sliding, and the area formed by the target size is used as the final denoising area.
8. The intelligent detection method for cell distribution of stem cell-oriented storage according to claim 7, wherein determining the window size adjustment value of the sliding according to the first window variation parameter and the second window variation parameter of the sliding comprises:
The product of the first window variable and the second window variable of the sliding is calculated, and then the value obtained by adding the product of the two window variable and the first window variable is used as the window size adjustment value of the sliding.
9. The intelligent detection method for cell distribution of stem cell storage according to claim 1, wherein the analysis of cell distribution according to the denoised stem cell image to obtain the distribution of stem cells to be detected comprises:
Performing edge detection on the denoised stem cell image to obtain the edges of each cell; connecting and repairing the edges of each cell to obtain the boundaries of each stem cell; and identifying and positioning the boundaries of each stem cell to obtain the distribution condition of the stem cells to be detected.
10. A stem cell storage-oriented cell distribution intelligent detection system, comprising a processor and a memory, wherein the processor is configured to process instructions stored in the memory to implement a stem cell storage-oriented cell distribution intelligent detection method according to any one of claims 1-9.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410443111.3A CN118134802A (en) | 2024-04-12 | 2024-04-12 | Stem cell storage-oriented cell distribution intelligent detection method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410443111.3A CN118134802A (en) | 2024-04-12 | 2024-04-12 | Stem cell storage-oriented cell distribution intelligent detection method and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN118134802A true CN118134802A (en) | 2024-06-04 |
Family
ID=91235971
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410443111.3A Pending CN118134802A (en) | 2024-04-12 | 2024-04-12 | Stem cell storage-oriented cell distribution intelligent detection method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN118134802A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN119693328A (en) * | 2024-12-03 | 2025-03-25 | 河北三臧生物科技有限公司 | A method for detecting density of induced pluripotent stem cells, computer equipment and medium |
-
2024
- 2024-04-12 CN CN202410443111.3A patent/CN118134802A/en active Pending
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN119693328A (en) * | 2024-12-03 | 2025-03-25 | 河北三臧生物科技有限公司 | A method for detecting density of induced pluripotent stem cells, computer equipment and medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113592861B (en) | Bridge crack detection method based on dynamic threshold | |
| CN109242853B (en) | An intelligent detection method for PCB defects based on image processing | |
| CN114937055A (en) | Image self-adaptive segmentation method and system based on artificial intelligence | |
| CN115908415B (en) | Edge-based defect detection method, device, equipment and storage medium | |
| CN117764864B (en) | MRI tumor visual detection method based on image denoising | |
| CN117557820B (en) | Quantum dot optical film damage detection method and system based on machine vision | |
| CN117974627A (en) | Surface crack detection method based on image processing | |
| CN116818778B (en) | Rapid and intelligent detection method and system for automobile parts | |
| CN117705815A (en) | Printing defect detection method based on machine vision | |
| CN115841434A (en) | Infrared image enhancement method for gas concentration analysis | |
| CN116416268B (en) | Method and device for detecting edge position of lithium battery pole piece based on recursion dichotomy | |
| CN115984148B (en) | Denoising enhancement method for high-throughput gene sequencing data | |
| CN117974706B (en) | A concave point segmentation method for rock slices based on dynamic threshold and local search | |
| CN112669295A (en) | Lithium battery pole piece defect detection method based on secondary threshold segmentation theory | |
| CN117541582B (en) | IGBT insulation quality detection method for high-frequency converter | |
| CN112308854A (en) | Automatic detection method and system for chip surface flaws and electronic equipment | |
| CN117437238B (en) | A visual inspection method for packaged IC surface defects | |
| CN116740728B (en) | Dynamic acquisition method and system for wafer code reader | |
| CN119151886B (en) | A PCBA board surface defect detection method based on computer vision | |
| CN118134802A (en) | Stem cell storage-oriented cell distribution intelligent detection method and system | |
| CN115330796A (en) | Copper wire tinning defect identification method | |
| CN117911419A (en) | Method and device for detecting steel rotation angle enhancement of medium plate, medium and equipment | |
| CN119477709A (en) | Fluorescence in situ hybridization image edge enhancement method based on illumination fusion and local binary | |
| CN116883401B (en) | Industrial product production quality detection system | |
| CN117218130A (en) | Image dimension reduction acquisition method and system for positioning iron element segregation defect |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |