Method for confirming cell mutation degree in TCT detection mode

Abstract

The invention discloses a method for confirming the cell mutation degree in a TCT detection mode. The method comprises the following steps: 1) preprocessing an original TCT cell staining image; 2) dividing pixels of the obtained grey-scale map into three categories of cell nucleuses, cell bodies and backgrounds, and respectively completing corresponding loop iterations; 3) performing segmentation, post-processing and closed operation in sequence from the obtained result maps of the three clusters, and determining a cell nucleus region result; 4) sequentially carrying out segmentation and post-processing operation in combination with the obtained cell nucleus region from the three obtained clustering result maps, and determining a cell body region result; and 5) according to the obtained cell nucleus region result and the obtained cell body region result, obtaining the nucleus-plasma ratio rho of the cells, and finally obtaining the characterization conclusion of the cell mutation degree. The method belongs to the technical field of images, and solves the problem that the cell mutation degree cannot be accurately determined due to the fact that the nucleus-plasma ratio of the cells is difficult to accurately calculate in the prior art.

Description

Technical Field

[0001] The present invention belongs to the field of image technology, is used for cell morphology analysis in the biomedical field, and relates to a method for confirming the degree of cell mutation under the TCT detection mode. Background Art

[0002] With the rapid development of biomedical imaging technology, automated image analysis has become a key research area in cell morphology analysis. Existing cell analysis methods often rely on manual experience, which is time-consuming and susceptible to human interference. This is particularly inefficient and prone to large errors when processing large numbers of cell images. Cell morphology is complex and variable, making existing methods unable to meet the demand for efficient and accurate large-scale cell image analysis.

[0003] The nuclear-to-cytoplasm ratio refers to the volume ratio of the cell nucleus to the cytoplasm, and is of great significance in the fields of cell biology, pathology, etc. As an important morphological indicator of cells, the nuclear-to-cytoplasm ratio directly affects the biological functions of cells, such as metabolism, proliferation, and differentiation. Currently, cell image processing technology has made certain progress in cell segmentation and recognition, but there are still deficiencies in the calculation of the nuclear-to-cytoplasm ratio and the extraction of morphological features. Although existing methods can effectively separate cells from the background, when calculating cell morphological features such as the nuclear-to-cytoplasm ratio, they still face challenges in segmentation accuracy and cell morphological diversity. Therefore, how to accurately calculate the nuclear-to-cytoplasm ratio and extract other morphological features has become a difficult problem in cell image analysis. Summary of the Invention

[0004] The purpose of the present invention is to provide a method for confirming the degree of cell mutation under TCT detection mode, which solves the problem that the existing technology is difficult to accurately calculate the nuclear-cytoplasmic ratio, making it impossible to accurately determine the degree of cell mutation.

[0005] The technical solution adopted by the present invention is a method for confirming the degree of cell mutation under TCT detection mode, which is implemented according to the following steps: Step 1: preprocessing the original TCT cell staining image; Step 2: The grayscale image obtained in step 1 The pixels are divided into three categories: cell nucleus, cell body, and background, and the corresponding loop iterations are completed respectively; Step 3: From the three clustering result images obtained in step 2, segmentation, post-processing, and closing operations are performed in sequence to determine the cell nucleus region results; Step 4: Based on the three clustering result images obtained in step 2, combined with the cell nucleus region obtained in step 3, segmentation and post-processing operations are performed in sequence to determine the cell body region result; Step 5: Based on the results of the cell nucleus area obtained in step 3 and the cell body area obtained in step 4, the nuclear-cytoplasmic ratio of the cell is obtained. ρ, and finally obtain the conclusion of the degree of cell mutation.

[0006] The beneficial effect of the present invention is that, through image segmentation and morphological closing operation processing, accurate segmentation and positioning of the cell nucleus area and the cell body area are achieved, effectively overcoming the shortcomings of existing methods in complex morphological cell analysis, and significantly improving the analysis efficiency while improving the accuracy of cell morphology analysis. In particular, it has significant advantages in the accuracy of large-scale cell image data processing and nuclear-cytoplasm ratio calculation, and can be widely used in cell biology research, disease diagnosis, drug screening and other fields. BRIEF DESCRIPTION OF THE DRAWINGS

[0007] Figure 1 is a sample image to be processed in embodiment 1 of the present invention; Figure 2 is a grayscale image obtained by the method of the present invention based on the dyeing attribute; Figure 3 It is a binary image of the candidate region of the cell nucleus obtained by the method of the present invention through three types of clustering; Figure 4 Yes Figure 3 The binary image of the candidate region of the cell nucleus is completed after the interference is removed; Figure 5 Yes Figure 4 The binary image of the cell nucleus region obtained after the closing operation; Figure 6 is a binary image of the cell body candidate region of the present invention; Figure 7 Yes Figure 6 The binary image of the cell body area after interference removal; Figure 8 is a sample image to be processed in embodiment 2 of the present invention; Figure 9 is a sample image to be processed in embodiment 3 of the present invention; Figure 10 is a sample image to be processed in embodiment 4 of the present invention; Figure 11 is a sample image to be processed in embodiment 5 of the present invention; Figure 12 This is the sample image that needs to be processed in Example 6 of the present invention. DETAILED DESCRIPTION

[0008] The present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0009] The method of the present invention, in TCT detection mode, aims to provide more efficient and accurate cell morphology analysis by automatically selecting a color channel suitable for segmentation, analyzing the characteristics of the cell nucleus and cell body in the selected color channel, and obtaining the nucleocytoplasm ratio value of the cell. The method is implemented according to the following steps: Step 1: Preprocess the original TCT cell staining image. In order to adapt to different dyeing conditions, the distribution of different color channels is statistically analyzed and one of the color channels is selected as the grayscale image for subsequent processing. The specific process is: Assume that the input sample image is of size The three-color channel TCT cell staining image is denoted as , It's the red channel. It's a green channel. is the blue channel; follow the steps below to determine a color channel and convert it into a grayscale image, recorded as , 1.1) Calculate the grayscale distribution of the three color channels of the stained image and obtain the mean value of each color and standard deviation , , the expression is: (1) (2) 1.2) Normalize the result obtained in step 1.1) and the expression is: , and standard deviation , (3) 1.3) According to the calculation results of formula (4) and (5), determine an optimal color channel as the grayscale image for subsequent processing, which is recorded as , the expression is: (4) (5) in, , .

[0010] Figure 1 This is the TCT cell staining image (i.e., sample image) to be processed in Example 1. After the operation in step 1, it is determined that Figure 2 The grayscale image shown is taken as the optimal grayscale image.

[0011] Step 2: Because the best grayscale image selected in step 1 includes other interference parts in addition to cells, the grayscale image obtained in step 1 is converted to The pixels are divided into three categories: cell nucleus, cell body, and background, and the corresponding loop iterations are completed respectively. The specific process is: 2.1) Calculate the number of pixels of different gray levels in the grayscale image , , the expression is: , (6) in, ,function Count the pixels that meet the conditions; 2.2) Calculate the class boundary thresholds of the three categories of cell nucleus, cell body, and background in the grayscale image , the expression is: (7) (8) in, is the ratio of the size of the cell nucleus in the grayscale image, is the ratio of the size of the cell body in the grayscale image; in this step, it is preferred , ; 2.3) Calculate the initial cluster centers of the three categories of cell nucleus, cell body and background, which are respectively denoted as , the expressions are: (9) (10) (11) 2.4) Solving the grayscale image The distance between all pixels and the three cluster centers, and at the same time, solve the grayscale image The minimum distances between all pixels and the three cluster centers are expressed as follows: (12) (13) in, , , ; 2.5) Let the sets of three categories, cell nucleus, cell body and background, be , solve the grayscale image The category of all pixels is expressed as: (14) 2.6) Update the three cluster centers. The expressions are: (15) (16) (17) in, Corresponding to The number of pixels in ; 2.7) Go to step 2.4) and iterate until Until it remains unchanged.

[0012] Step 3: From the three clustering results obtained in step 2, perform segmentation, post-processing, and closing operations in sequence to determine the cell nucleus area results, which are recorded as binary images. The specific process is: 3.1) Segmentation to obtain a binary image of the candidate nucleus region , the expression is: (18) in, , , 3.2) Considering that the candidate region of the cell nucleus includes some interference, it is necessary to post-process it and convert the binary image obtained by formula (18) into Perform labeling (this is the existing technology) and take the connected domain with the largest area, which is recorded as ; 3.3) Candidate regions of cell nuclei after removing interference , select the size , the structural element with the origin at the center point, performs a closing operation to obtain the cell nucleus area, and uses a binary image means that the expression is: (19) in, , , For experience value, preferred , that is, the cell nuclear area was obtained.

[0013] Step 4: From the three clustering result images obtained in step 2, combined with the cell nucleus region obtained in step 3, perform segmentation and post-processing operations in sequence to determine the cell body region result, and use the binary image The specific process is: 4.1) Segmentation to obtain a binary image of the cell body candidate region , the expression is: (20) in, , ; 4.2) Consider the candidate cell body regions Including some interference, so it needs to be post-processed and the binary image calculated by formula (20) Perform labeling (existing technology) and take the largest connected domain to obtain the cell body region result, which is recorded as a binary image. .

[0014] Step 5: Based on the results of the cell nucleus area obtained in step 3 and the cell body area obtained in step 4, the nuclear-cytoplasmic ratio of the cell is obtained. ρ , the expression is: (twenty one) From this point on, the degree of cell mutation can be determined based on the nuclear-to-cytoplasmic ratio. Generally speaking, a larger nuclear-to-cytoplasmic ratio indicates a greater degree of cell mutation, ensuring the accuracy and efficiency of cell morphology analysis.

[0015] Example 1 For example Figure 1 The sample image shown was stained with lighter color and shows more texture of the cell body.

[0016] This embodiment 1 is specifically implemented according to the following steps: Step 1: Preprocess the original TCT cell staining image. In order to adapt to different dyeing conditions, the distribution of different color channels is statistically analyzed and one of the color channels is selected as the grayscale image for subsequent processing. The specific process is: Assume that the input sample image is of size The three-color channel TCT cell staining image is denoted as , It's the red channel. It's a green channel. is the blue channel; follow the steps below to determine a color channel and convert it into a grayscale image, recorded as , 1.1) Calculate the grayscale distribution of the three color channels of the stained image and obtain the mean value of each color and standard deviation , , the expression is: (1) (2) 1.2) Normalize the result obtained in step 1.1) and the expression is: , and standard deviation , (3) 1.3) According to the calculation results of formula (4) and (5), determine an optimal color channel as the grayscale image for subsequent processing, which is recorded as , the expression is: (4) (5) in, , .

[0017] Through the above steps, the normalized standard deviations of the color channels calculated are: 16.50, 15.53 and 15.63, so the first channel is selected as the grayscale image for calculating the nuclear-cytoplasmic ratio.

[0018] Figure 1 This is the TCT cell staining image (i.e., sample image) to be processed in Example 1. After the operation in step 1, Figure 2 Grayscale image shown.

[0019] Step 2: Because the grayscale image obtained in step 1 includes other interference parts in addition to cells, the grayscale image obtained in step 1 is converted to The pixels are divided into three categories: cell nucleus, cell body, and background, and the corresponding loop iterations are completed respectively. The specific process is: 2.1) Calculate the number of pixels of different gray levels in the grayscale image , , the expression is: , (6) in, ,function Count the pixels that meet the conditions; 2.2) Calculate the class boundary thresholds of the three categories of cell nucleus, cell body, and background in the grayscale image , the expression is: (7) (8) in, is the ratio of the size of the cell nucleus in the grayscale image, is the ratio of the size of the cell body in the grayscale image; in this step, the parameter value is , ; 2.3) Calculate the initial cluster centers of the three categories of cell nucleus, cell body and background, which are respectively denoted as , the expressions are: (9) (10) (11) 2.4) Solving the grayscale image The distance between all pixels and the three cluster centers, and at the same time, solve the grayscale image The minimum distances between all pixels and the three cluster centers are expressed as follows: (12) (13) in, , , ; 2.5) Let the sets of three categories, cell nucleus, cell body and background, be , solve the grayscale image The category of all pixels is: (14) 2.6) Update the three cluster centers. The expressions are: (15) (16) (17) in, Corresponding to The number of pixels in ; 2.7) Go to step 2.4) and iterate until Until it remains unchanged.

[0020] Step 3: From the three clustering results obtained in step 2, perform segmentation, post-processing, and closing operations in sequence to determine the cell nucleus area results, which are recorded as binary images. The specific process is: 3.1) Segmentation to obtain a binary image of the candidate nucleus region , the expression is: (18) in, , , In Example 1, according to formula (18), Figure 3 The results shown; 3.2) Considering that the candidate region of the cell nucleus includes some interference, it is necessary to post-process it and convert the binary image obtained by formula (18) into Perform labeling (this is the existing technology) and take the connected domain with the largest area, which is recorded as , and get Figure 4 The results shown; 3.3) Candidate regions of cell nuclei after removing interference , select the size , the structural element with the origin at the center point, performs a closing operation to obtain the cell nucleus area, and uses a binary image means that the expression is: (19) in, , , The nucleus area was obtained when the value was 7.

[0021] In Example 1, Figure 4 After closing operation, we get Figure 5 Results for the nuclear region are shown; Step 4: From the three clustering result images obtained in step 2, combined with the cell nucleus region obtained in step 3, perform segmentation and post-processing operations in sequence to determine the cell body region result, and use the binary image The specific process is: 4.1) Segmentation to obtain a binary image of the cell body candidate region , the expression is: (20) in, , , In Example 1, according to formula (20), we get Figure 6 The results shown; 4.2) Consider the candidate cell body regions Including some interference, so it needs to be post-processed and the binary image calculated by formula (20) Perform labeling (existing technology) and take the largest connected domain to obtain the cell body region result, which is recorded as a binary image. In Example 1, Figure 7 shown.

[0022] Step 5: Based on the results of the cell nucleus area obtained in step 3 and the cell body area obtained in step 4, the nuclear-cytoplasmic ratio of the cell is obtained. ρ , the expression is: (twenty one) Through the processing of each step, we can get Figure 2-Figure 7 The final nuclear-cytoplasmic ratio was 0.15:1, which indicates that the degree of cell mutation is not large.

[0023] Example 2 According to the steps of the above-mentioned method of the present invention, Figure 8 The sample image shown in the figure has a strong interference area above the cell body. The color of the strong interference area is different from that of the cell body. The cell body has more texture. The normalized standard deviations of the calculated color channels are 14.13, 13.57, and 13.70, respectively. Therefore, the first channel is selected as the grayscale image for calculating the nuclear-cytoplasmic ratio. is 7, , The final nuclear-cytoplasmic ratio was 0.27:1, which indicates that the degree of cell mutation is not great.

[0024] Example 3 According to the steps of the above-mentioned method of the present invention, Figure 9 In the sample image shown, the staining condition is that the cell body is partially overlapped by the plasma of another cell, forming a strong interference. The normalized standard deviations of the calculated color channels are: 22.87, 24.38 and 26.40 respectively. Therefore, the third channel is selected as the grayscale image for calculating the nuclear-cytoplasmic ratio. is 9, parameter , The final nuclear-cytoplasmic ratio was 0.31:1, which indicates that the degree of cell mutation is not great.

[0025] Example 4 According to the steps of the above-mentioned method of the present invention, Figure 10 In the sample image shown, the staining condition is that the cell body is partially overlapped by the plasma of another cell, forming a strong interference. The normalized standard deviations of the calculated color channels are: 11.60, 13.13 and 14.43 respectively. Therefore, the third channel is selected as the grayscale image for calculating the nuclear-cytoplasmic ratio. is 7, parameter , The final nuclear-cytoplasmic ratio was 0.75:1, which indicates that the degree of cell mutation is relatively high.

[0026] Example 5 According to the steps of the above-mentioned method of the present invention, Figure 11 In the sample image shown, the cell body is unusually elongated. The normalized standard deviations of the color channels calculated are 11.02, 12.80, and 12.16, respectively. Therefore, the second channel is selected as the grayscale image for calculating the nuclear-cytoplasmic ratio. is 11, parameter , The final nuclear-cytoplasmic ratio was 0.62:1, which indicates that the degree of cell mutation is relatively high.

[0027] Example 6 According to the steps of the above-mentioned method of the present invention, Figure 12 In the sample image shown, the cell body is severely unevenly stained. The normalized standard deviations of the color channels are calculated to be 18.56, 21.29, and 20.09, respectively. Therefore, the second channel is selected as the grayscale image for calculating the nuclear-cytoplasmic ratio. is 13, parameter , The final nuclear-cytoplasmic ratio was 0.24:1, which indicates that the degree of cell mutation is not large.

Claims

1. A method for confirming the degree of cell mutation in TCT detection mode, characterized in that: Follow these steps to implement: Step 1: preprocessing the original TCT cell staining image; Step 2: The grayscale image obtained in step 1 The pixels are divided into three categories: cell nucleus, cell body, and background, and the corresponding loop iterations are completed respectively; Step 3: From the three clustering result images obtained in step 2, segmentation, post-processing, and closing operations are performed in sequence to determine the cell nucleus region results; Step 4: Based on the three clustering result images obtained in step 2, combined with the cell nucleus region obtained in step 3, segmentation and post-processing operations are performed in sequence to determine the cell body region result; Step 5: Based on the results of the cell nucleus area obtained in step 3 and the cell body area obtained in step 4, the nuclear-cytoplasmic ratio of the cell is obtained. ρ , and finally obtain the conclusion of the degree of cell mutation.

2. The method for confirming the degree of cell mutation under TCT detection mode according to claim 1, characterized in that: In step 1, the specific process is: Assume that the input sample image is of size The three-color channel TCT cell staining image is denoted as , It's the red channel. It's a green channel. is the blue channel; Follow the steps below to determine a color channel and convert it into a grayscale image, denoted as , 1.1) Calculate the grayscale distribution of the three color channels of the stained image and obtain the mean value of each color. and standard deviation , , the expression is: （1） （2） 1.2) Normalize the result obtained in step 1.1) and the expression is: , and standard deviation , (3) 1.3) According to the calculation results of formula (4) and (5), determine an optimal color channel as the grayscale image for subsequent processing, which is recorded as , the expression is: （4） （5） in, , .

3. The method for confirming the degree of cell mutation under TCT detection mode according to claim 1, characterized in that: In step 2, the specific process is: 2.1) Calculate the number of pixels of different gray levels in the grayscale image , , the expression is: ，（6） in, ,function Count the pixels that meet the conditions; 2.2) Calculate the class boundary thresholds of the three categories of cell nucleus, cell body, and background in the grayscale image , the expression is: （7） （8） in, is the ratio of the size of the cell nucleus in the grayscale image, is the ratio of the size of the cell body in the grayscale image; 2.3) Calculate the initial cluster centers of the three categories of cell nucleus, cell body and background, which are respectively denoted as , the expressions are: （9） （10） （11） 2.4) Solving the grayscale image The distance between all pixels and the three cluster centers, and at the same time, solve the grayscale image The minimum distances between all pixels and the three cluster centers are expressed as follows: （12） （13） in, , , ; 2.5) Let the sets of three categories, cell nucleus, cell body and background, be , solve the grayscale image The category of all pixels is: （14） 2.6) Update the three cluster centers. The expressions are: （15） （16） （17） in, Corresponding to The number of pixels in ; 2.7) Go to step 2.4) and iterate until Until it remains unchanged.

4. The method for confirming the degree of cell mutation under TCT detection mode according to claim 3, characterized in that: In step 2.2), , .

5. The method for confirming the degree of cell mutation under TCT detection mode according to claim 1, characterized in that: In step 3, the specific process is: 3.1) Segmentation to obtain a binary image of the candidate nucleus region , the expression is: （18） in, , ; 3.2) Perform post-processing and convert the binary image obtained by formula (18) Perform labeling and take the largest connected domain, recorded as ; 3.3) Candidate regions of cell nuclei after removing interference , select the size , the structural element with the origin at the center point, performs a closing operation to obtain the cell nucleus area, and uses a binary image means that the expression is: （19） in, , , is an empirical value, that is, the cell nucleus area is obtained.

6. The method for confirming the degree of cell mutation under TCT detection mode according to claim 5, characterized in that: In step 3.3), .

7. The method for confirming the degree of cell mutation under TCT detection mode according to claim 1, characterized in that: In step 4, the specific process is: 4.1) Segmentation to obtain a binary image of the cell body candidate region , the expression is: （20） in, , ; 4.2) Perform post-processing to convert the binary image Perform labeling and take the largest connected domain to obtain the cell body region result, which is recorded as a binary image. .

8. The method for confirming the degree of cell mutation under TCT detection mode according to claim 1, characterized in that: In step 5, the nuclear-cytoplasmic ratio ρ The expression is: .

Images