A method and system for calculating cell wall grayscale, hollowness and thickness

Abstract

This invention discloses a method and system for calculating cell wall grayscale, hollowing ratio, and thickness. This method, which involves the intersection of computer vision and biotechnology, involves preprocessing fungal cell images taken under a microscope and generating cell wall location labels using semi-automatic annotation. A target segmentation model is built and trained based on the preprocessed images, and model parameters are optimized using a loss function to segment and extract cell wall contours. Based on the segmentation results, the cell wall grayscale mean, hollowing ratio, and thickness are calculated, and the model's generalization capability is enhanced through data enhancement and algorithm optimization. The method can accurately segment cell walls in images and accurately calculate cell wall-related parameters, with a high degree of automation and significant reductions in labor and time costs.

Description

Technical Field

[0001] The present invention relates to the interdisciplinary field of computer vision technology and biotechnology, and specifically to a method and system for calculating the grayscale, hollowness, and thickness of a cell wall. Background Art

[0002] The study of cells is one of the most common areas of biological research today, and comparing the ability of different enzymes to degrade cell walls is a crucial area of ​​research. As a crucial barrier protecting cells, the composition and structure of the cell wall determine the cell's adaptability and resistance to the external environment. Enzymes can selectively degrade certain key components of the cell wall. Therefore, this research is of great significance to biology. Currently, determining the ability of different enzymes to degrade cell walls often relies on manual detection methods. However, manual observation is subjective, variable, and time-consuming, making it inefficient and inefficient, making it difficult to meet the needs of current cell wall research.

[0003] Methods for automated image processing using computers include algorithms based on pixel-level threshold segmentation, filter box threshold segmentation, and Otsu threshold segmentation. These algorithms have the problems of low recall rate, poor noise resistance, and high requirements for image quality. Pre-prepared template materials are used for matching and comparison with images. Matching methods include color matching and variance matching. These methods have the problems of low recall rate, strong data dependence, and long time consumption. Feature descriptor detection methods such as Harris corner detection and its derivative detection methods, SIFT operator, etc. are used to construct image features of the detection target. During detection, the target is detected through a classification model of image features. This method has the problems of being unable to adapt to size changes and having a high false detection rate. Machine learning methods used for image classification and target detection are interpretable but have low accuracy. Deep learning has improved the accuracy, but the model is a black box and difficult to interpret. Even if the target object is detected, it is difficult to divide the foreground and background in a fine-grained manner. Summary of the Invention

[0004] In view of the above-mentioned problems, the present invention is proposed.

[0005] Therefore, the technical problems solved by the present invention are: the existing manual detection methods are not accurate enough in identifying the ability of different enzymes to decompose cell walls, the understanding of cell walls is vague when manually segmenting cell walls, the existing detection methods have a low degree of automation, a high degree of manual participation, and are time-consuming, making it difficult to meet experimental needs, and how to scientifically compare the ability of different enzymes to decompose cell walls based on accurately calculated parameters.

[0006] To solve the above technical problems, the present invention provides the following technical solutions: a method for calculating cell wall grayscale, hollowing ratio, and thickness is based on preprocessing of fungal cell images taken under a microscope, and cell wall position labels are generated by semi-automatic annotation; a target segmentation model is built and trained based on the preprocessed images, and the model parameters are optimized through a loss function to segment and extract the cell wall contour; the cell wall grayscale mean, hollowing ratio, and thickness are calculated based on the segmentation results, and the model generalization ability is improved through data enhancement and algorithm optimization; the loss function includes a cross-entropy loss function and an auxiliary loss function, the cross-entropy loss function is used to compare the difference between the output and the actual label, and evaluate the prediction accuracy of the model at the pixel level, and the auxiliary loss function is used to measure the similarity between the model prediction output and the actual label, and determine the degree of prediction error.

[0007] As a preferred solution of the method for calculating cell wall grayscale, hollowness, and thickness described in the present invention, the preprocessing includes reading ruler information, manually measuring the ruler information, and calculating the actual length represented by the pixel in combination with the image pixel width for physical unit conversion of the cell wall thickness parameter.

[0008] As a preferred solution of the method for calculating cell wall grayscale, hollowness and thickness described in the present invention, the target segmentation model includes adopting the YOLO11 segmentation model. YOLO11 consists of a backbone network, a neck network and a head network. The backbone network is used to extract image features, the neck network is used to further process and fuse image features, and the head network is used for target monitoring and classification. The target segmentation model is used to perform image segmentation and identify fungal cell walls, and extract cell wall contours.

[0009] As a preferred solution of the method for calculating cell wall grayscale, hollowness and thickness described in the present invention, the data enhancement includes rotation, flipping and scaling;

[0010] Rotation, by rotating the image at random angles within a given parameter range to simulate shooting at different angles;

[0011] Flip, by specifying the probability parameter of flip, the probability of flipping is controlled. The flips used include horizontal flipping and vertical flipping.

[0012] Scaling: Given a parameter, the image is randomly scaled within the parameter range to simulate the performance of the target at different distances.

[0013] As a preferred solution of the method for calculating cell wall grayscale, hollowness and thickness described in the present invention, the cell wall grayscale mean includes extracting the grayscale sum and the total number of pixels in the cell wall area through a mask, calculating the grayscale value, and taking the average value to reflect the overall brightness characteristics of the cell wall.

[0014] As a preferred solution of the method for calculating the grayscale, hollowness and thickness of the cell wall described in the present invention, the hollowness ratio includes extracting the grayscale value of each pixel in the cell wall area through a mask, determining that the pixel is white, calculating the total number of white pixels and the total number of pixels, and calculating the hollowness ratio to characterize the internal integrity and degree of damage of the cell wall.

[0015] As a preferred solution of the method for calculating cell wall grayscale, hollowness and thickness described in the present invention, the thickness includes: segmenting the cell wall from the cell through the target segmentation model to obtain a set of edge pixels, selecting corner points, determining the effective corner points by calculating the distance between the corner points and the edge of the image, using the KDTree algorithm to calculate the shortest distance between the two parts of pixel points, and taking the median of the distance as the cell wall thickness.

[0016] As a preferred solution of the method for calculating cell wall grayscale, hollowness and thickness described in the present invention, the corner points include selecting specific pixel points near the edge of the cell image as the key reference for calculating the cell wall thickness, determining the image edge and taking the average value as the target edge, traversing the pixel point set, and calculating the distance between each pixel point and the target edge.

[0017] Corner point selection includes the points closest to the left, right, top, and bottom edges of the image. By calculating the distance from the point to the edge, valid corner points are screened out.

[0018] As a preferred solution of the method for calculating cell wall grayscale, hollowness, and thickness described in the present invention, the algorithm optimization includes using a visual evaluation method based on the CDF curve to verify the effectiveness and rationality of the algorithm by comparing and analyzing the parameter changes before and after the cell wall is decomposed by the enzyme.

[0019] Another object of the present invention is to provide a cell wall grayscale, hollowness, and thickness calculation system, which can build and train a target segmentation model by preprocessing images, optimize model parameters through a loss function, segment and extract cell wall contours, and solve the problem of low accuracy contained in the current manual detection and calculation of cell wall parameter technologies.

[0020] As a preferred solution of the cell wall grayscale, hollowness and thickness calculation system described in the present invention, it includes: an image preprocessing module, a target segmentation module, and a parameter calculation module. The image preprocessing module is used to annotate the input cell image and calculate the scale information to generate the cell wall position label; the target segmentation module is used to build and train the target segmentation algorithm model, optimize the image segmentation loss function, and extract the cell wall contour; the parameter calculation module is used to calculate the grayscale mean, hollowness ratio and thickness of the cell wall based on the segmented cell image, and perform data enhancement and algorithm optimization.

[0021] A computer device includes a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement a method for calculating the grayscale, hollowness and thickness of a cell wall.

[0022] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of a method for calculating the grayscale, hollowness and thickness of a cell wall.

[0023] Beneficial effects of the present invention: A method for calculating cell wall grayscale, hollowing ratio and thickness provided by the present invention is based on preprocessing of fungal cell images taken under a microscope, and a semi-automatic labeling method is used to generate cell wall position labels, combining automatic labeling with manual verification to ensure the high quality of label files while improving efficiency. A target segmentation model is built and trained based on the preprocessed image, and the model parameters are optimized through a loss function. The cell wall contour is segmented and extracted, providing a reliable framework for future use and ensuring the accuracy of the cell wall parameters calculated by the algorithm. The cell wall grayscale mean, hollowing ratio and thickness are calculated based on the segmentation results, and the model generalization ability is improved through data enhancement and algorithm optimization, providing reliable data support for biological experiments, with a high degree of automation and reduced labor and time costs. The present invention achieves better results in terms of the accuracy of identifying the cell wall decomposition ability of different enzymes, the understanding of the cell wall and the automation of the detection method. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0025] Figure 1 This is an overall flow chart of a method for calculating cell wall grayscale, hollowness, and thickness provided in the first embodiment of the present invention.

[0026] Figure 2 A healthy cell collection image is provided for a method for calculating cell wall grayscale, hollowness, and thickness according to the second embodiment of the present invention.

[0027] Figure 3 The second embodiment of the present invention provides a method for calculating the grayscale, hollowness and thickness of a cell wall, which includes images of cells decomposed by enzymes.

[0028] Figure 4 This is the segmentation result of a method for calculating cell wall grayscale, hollowness, and thickness provided by the second embodiment of the present invention.

[0029] Figure 5 This is a mask coverage result diagram of a method for calculating cell wall grayscale, hollowness, and thickness provided by the second embodiment of the present invention.

[0030] Figure 6 Figure a shows an example of calculation results of a method for calculating cell wall grayscale, hollowness, and thickness provided in the second embodiment of the present invention.

[0031] Figure 7 Figure b is an example of the calculation results of a method for calculating cell wall grayscale, hollowness, and thickness provided in the second embodiment of the present invention.

[0032] Figure 8 Figure c is an example of the calculation results of a method for calculating cell wall grayscale, hollowness, and thickness provided in the second embodiment of the present invention.

[0033] Figure 9 Figure a is a CDF graph of the calculation results of a method for calculating cell wall grayscale, hollowness, and thickness provided in the second embodiment of the present invention.

[0034] Figure 10 Figure b is a CDF graph of the calculation results of a method for calculating cell wall grayscale, hollowness, and thickness provided in the second embodiment of the present invention.

[0035] Figure 11 Figure c is a CDF graph of the calculation results of a method for calculating cell wall grayscale, hollowness, and thickness provided by the second embodiment of the present invention. DETAILED DESCRIPTION

[0036] To make the above-mentioned objects, features, and advantages of the present invention more clearly understood, the following detailed description of the specific embodiments of the present invention is given in conjunction with the accompanying drawings. It is obvious that the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in this field without creative work should fall within the scope of protection of the present invention.

[0037] Example 1, reference Figure 1 , as one embodiment of the present invention, provides a method for calculating cell wall grayscale, hollowness, and thickness, including:

[0038] S1: Preprocess fungal cell images taken under a microscope and generate cell wall position labels using a semi-automatic annotation method.

[0039] Furthermore, the preprocessing includes reading the ruler information, manually measuring the ruler information, and calculating the real length represented by the pixel in combination with the image pixel width for the physical unit conversion of the cell wall thickness parameter.

[0040] It should be noted that a preferred solution for the scale information specifically includes manually measuring the unit scale and image length data, combining the image pixel width to obtain the pixel length occupied by the scale, and then calculating the actual length represented by one pixel from the actual length symbolized by the scale.

[0041] It should also be noted that by reading the microscope image scale information and manually measuring the scale length, the actual physical size represented by each pixel is calculated in combination with the image pixel width, thereby achieving accurate conversion of image physical units and improving the accuracy and reliability of quantitative analysis of cell morphology.

[0042] S2: Build and train the target segmentation model based on the preprocessed image, optimize the model parameters through the loss function, and segment and extract the cell wall contour.

[0043] Furthermore, the target segmentation model includes the use of the YOLO11 segmentation model. YOLO11 consists of a backbone network, a neck network, and a head network. The backbone network is used to extract image features, the neck network is used to further process and fuse image features, and the head network is used for target monitoring and classification. The target segmentation model is used to perform image segmentation and identify fungal cell walls, and extract cell wall contours.

[0044] It should be noted that the use of loss functions to optimize model parameters includes using loss functions when performing target segmentation on images. The loss function of image segmentation consists of a cross entropy loss function and an auxiliary loss function, which can be expressed as:

[0045] ,

[0046] in, is the loss function, Cross entropy loss compares the difference between the output and the actual label to evaluate the prediction accuracy of the model at the pixel level; For auxiliary loss, the similarity between the model's predicted output and the actual label is measured to determine the extent of the prediction error.

[0047] The cross entropy loss function is expressed as:

[0048] ,

[0049] in, is the cross entropy loss function, Expressed as image width, Expressed as image height, Represents the probability of a pixel being predicted to be in a certain class.

[0050] The auxiliary loss function is expressed as:

[0051] ,

[0052] in, is the auxiliary loss function, Represented as the ground truth pixel classification for segmentation, Denote as the predicted pixel classification, Expressed as and The intersection of Indicates the parameter added to prevent the denominator from being zero.

[0053] It should also be noted that the YOLO11 segmentation model is used for feature extraction and target detection, and the loss function is combined to optimize the model parameters to achieve high-precision segmentation and contour extraction of fungal cell walls, solve the problem of inaccurate cell wall boundary recognition in complex microscope images by traditional segmentation models, improve the accuracy and precision of cell wall segmentation, and enhance the robustness of the model.

[0054] S3: Calculate the cell wall grayscale mean, hollowing ratio, and thickness based on the segmentation results, and improve the model generalization ability through data enhancement and algorithm optimization.

[0055] Furthermore, data augmentation includes rotation, flipping and scaling.

[0056] Rotation: By rotating the image at random angles within a given parameter range, we can simulate shooting at different angles and improve the model's robustness to changes in target pose.

[0057] Flipping, by specifying the probability parameter of flipping, controls the possibility of flipping and increases data diversity. The flipping used includes horizontal flipping and vertical flipping.

[0058] Scaling: Given a parameter, the image is randomly scaled within the parameter range to simulate the performance of the target at different distances, improve the model's robustness to changes in target size, and increase data diversity.

[0059] The rotation method is expressed as:

[0060] ,

[0061] ,

[0062] in, Represents a specified rotation angle of 180°. and Expressed as the pixel coordinates before and after the transformation.

[0063] The scaling method is expressed as:

[0064] ,

[0065] ,

[0066] in, Expressed as a scaled ratio (range 1± ), Represents the specified scaling factor of 0.2, and Expressed as the pixel coordinates before and after the transformation.

[0067] It should be noted that the cell wall grayscale mean includes the grayscale sum of the pixels in the cell wall area extracted by mask and the total number of pixels, and the grayscale value is calculated and averaged to reflect the overall brightness characteristics of the cell wall.

[0068] The mean gray value of the cell wall is expressed as:

[0069] ,

[0070] ,

[0071] ,

[0072] in, Represented as an image The pixel value at position (i, j), Represented as the value of the position in the mask map, Expressed as the sum of the gray levels of the image, Expressed as the sum of the pixels of the image, Expressed as the average gray value of the image.

[0073] It should be noted that the hollowing ratio includes extracting the grayscale value of each pixel in the cell wall area through a mask, determining the pixels that are white, calculating the total number of white pixels and the total number of pixels, and calculating the hollowing ratio to characterize the internal integrity and degree of damage of the cell wall.

[0074] The hollowing ratio is expressed as:

[0075] ,

[0076] ,

[0077] ,

[0078] in, Represented as the pixel value of image I at position (i, j), Represented as the value of the position in the mask map, It is represented by the grayscale value of each pixel in the mask, where pixels <125 are black and pixels >125 are white. Expressed as the sum of the pixels of the image, Expressed as the total number of white pixels, Expressed as the hollowing ratio of the image.

[0079] It should be noted that the thickness includes segmenting the cell wall from the cell through the target segmentation model to obtain a set of edge pixels, selecting corner points, determining the effective corner points by calculating the distance between the corner points and the edge of the image, and using the KDTree algorithm to calculate the shortest distance between the two parts of pixels, and taking the median of the distance as the cell wall thickness.

[0080] Thickness is expressed as:

[0081] ,

[0082] in, Expressed as cell wall thickness, Represented as from the beginning to the second dividing point, Expressed as from the second dividing point to the end, and Represented as two test corner points.

[0083] It should also be noted that the corner points include selecting specific pixel points near the edge of the cell image as the key reference for calculating the cell wall thickness, determining the image edge and taking the average value as the target edge, traversing the pixel point set, and calculating the distance between each pixel point and the target edge.

[0084] Corner point selection includes the points closest to the left, right, top, and bottom edges of the image. By calculating the distance from the point to the edge, valid corner points are screened out.

[0085] Corner points are represented as:

[0086] ,

[0087] ,

[0088] ,

[0089] ,

[0090] in, Represented as the left edge of the image, Represented as the right edge of the image, Represented as the upper edge of the image, Represented as the lower edge of the image, Expressed as the horizontal coordinate position of the pixel point, Expressed as the vertical coordinate position of the pixel point, Expressed as the distance from the pixel to the left edge of the image, Expressed as the distance from the pixel to the right edge of the image, Expressed as the distance from the pixel to the top edge of the image, Expressed as the distance in pixels from the bottom edge of the image.

[0091] It should also be noted that the algorithm optimization includes the use of a visual evaluation method based on the CDF curve to verify the effectiveness and rationality of the algorithm by comparing and analyzing the changes in parameters before and after the cell wall is decomposed by the enzyme.

[0092] The CDF curve visualization method evaluates the central tendency of the parameter data (such as median and quantile), the degree of dispersion (steepness of the curve), and the distribution shape (symmetry or skewness), focusing on the comparison of the curve position, curve trend, and the flatness of the curve tail. It is used to analyze the differences in parameter size, central tendency, and number of extreme values ​​before and after the cell wall is decomposed by the enzyme.

[0093] It should also be noted that the generalization ability of the model is improved through data enhancement, and the quantitative calculation of grayscale mean, hollowing ratio and thickness is combined to achieve accurate analysis of the morphological characteristics of fungal cell walls. The CDF curve is used to visualize the evaluation of algorithm optimization effects, solve the problem of low accuracy in measuring cell wall parameters in microscope images, enhance the adaptability of the model, and improve the quantification accuracy and reliability of cell wall structure.

[0094] Example 2, reference Figure 2-Figure 11 , which is an embodiment of the present invention, provides a method for calculating the grayscale, hollowness and thickness of the cell wall. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through economic benefit calculation and simulation experiments.

[0095] First, the Anylabeling software package was used to manually label 88 images taken with the microscope, 79 of which were used for training and 9 for validation. Manual labeling can improve the accuracy of the model, such as Figure 2 and Figure 3 As shown in the figure, the healthy cell collection image and the enzyme-decomposed cell collection image are collected. To expand the data volume, the image is enhanced, including rotating 180 degrees, horizontal / vertical flipping probability 0.25, and scaling ratio 0.2. After calculating the loss through the loss function, the target segmentation model parameters are updated through back propagation. After the target segmentation model is trained, as shown in the figure, Figure 4 The model shown is used to separate the cell wall from the cell.

[0096] The grayscale value statistics of the obtained cell wall area are performed based on the mask technology to obtain the total grayscale value and average value of the cell wall. During the processing, the mask area is first constructed based on the polygon outline information defined in the label file, and the non-cell wall part is removed using bit operations to retain only the target area. Secondly, the image is read in grayscale mode (cv2.IMREAD_GRAYSCALE) to ensure that the single-channel grayscale data of the image is not interfered with by redundant information. Then, as Figure 5 As shown in the following example, the np.sum() function is used to calculate the sum of the grayscale values ​​of all pixels in the mask area, and the np.count_nonzero() function is combined to obtain the total number of pixels in the area, as shown in the following example: Figure 6 The final calculated average grayscale value is shown as the mask coverage result. During execution, set the image reading mode to grayscale mode (single channel) and specify the mapping between image and label files. Specifically, the image file suffix is ​​".TIF" and the label file is ".json." Both file names must have the same prefix. The final calculation results are output to the file gray_value_sum_and_average2.txt, which records the total grayscale value and average for each image.

[0097] The hollowing ratio of the cell wall area was calculated by the threshold method, as Figure 7 The cell wall hollowing ratio results shown here measure the integrity of the cell wall structure and the extent of damage after enzyme disruption. A mask region is generated by reading polygon coordinates from the annotation file. Bitwise operations are used to extract the cell wall region. Within this region, the grayscale value is determined to be above a set threshold. Finally, the ratio of high-grayscale pixels to the total number of pixels in the region is calculated as the hollowing ratio. The calculation requires a grayscale threshold for high grayscale detection; pixels with grayscale values ​​greater than 125 are considered bright spots. Furthermore, image files and corresponding label files are automatically matched based on their filename prefixes. The default suffix for images is ".TIF" and the suffix for label files is ".json." The calculated results are output to a text file (hollow_out_ratio.txt) for subsequent analysis and comparison.

[0098] The thickness of the cell wall is quantitatively evaluated based on the segmented cell wall edge pixel set, such as Figure 8 The cell wall thickness calculation results shown are used to measure the physical structure changes under low enzyme treatment conditions. During the specific operation process, the program automatically extracts the corner points near the upper, lower, left, and right edges of the image from the contour points, and divides the path based on these corner points. Then, a nearest neighbor search structure is constructed using KDTree to query the closest distance between two edge contours point by point. Finally, the median of all distances is taken as the thickness estimate. Through the detailed experimental steps combined with the CDF curve graph, it can be seen that Figure 9The gray value of the enzymatically hydrolyzed cells is significantly higher than that of healthy cells. Figure 10 The hollowing ratio distribution curve of the enzymatically hydrolyzed cells is obviously shifted to the right, and the structure is damaged, as shown in Figure 11 Although the median thickness of the enzymatically hydrolyzed cells does not change significantly, the extension of the CDF curve tail shows an increase in discreteness, thus verifying the effectiveness of the present invention. In summary, the present invention can accurately segment the cell wall in the image and accurately calculate the cell wall-related parameters. At the same time, the present invention can provide reliable and accurate data support for the subsequent comparison of the decomposition ability of different enzymes on the cell wall, and solves the problem of unclear understanding of the cell wall during manual annotation. The segmentation and calculation process of the present invention is completed by the program itself, with a high degree of automation, which greatly reduces labor costs and time costs.

[0099] Example 3 is an embodiment of the present invention, which provides a cell wall grayscale, hollowness, and thickness calculation system, including an image preprocessing module, a target segmentation module, and a parameter calculation module.

[0100] Among them, the image preprocessing module is used to annotate the input cell image and calculate the scale information to generate the cell wall position label;

[0101] Among them, the target segmentation module is used to build and train the target segmentation algorithm model, optimize the loss function of image segmentation, and extract the cell wall contour;

[0102] Wherein: the parameter calculation module is used to calculate the grayscale mean, hollowing ratio and thickness of the cell wall based on the segmented cell image, and perform data enhancement and algorithm optimization.

Claims

1. A method for calculating cell wall grayscale, hollowness, and thickness, characterized in that: include: Based on the preprocessing of fungal cell images taken under a microscope, cell wall position labels were generated using a semi-automatic annotation method; Build and train the target segmentation model based on the preprocessed images, optimize the model parameters through the loss function, and segment and extract the cell wall contours; The cell wall grayscale mean, hollowing ratio, and thickness are calculated based on the segmentation results, and the model generalization ability is improved through data enhancement and algorithm optimization. The hollowing ratio includes extracting the gray value of each pixel in the cell wall area through a mask, determining the pixel is white, calculating the total number of white pixels and the total number of pixels, and calculating the hollowing ratio to characterize the integrity and damage degree of the cell wall; The thickness includes segmenting the cell wall from the cell using a target segmentation model to obtain a set of edge pixels, selecting corner points, determining valid corner points by calculating the distance between the corner points and the edge of the image, and using a KDTree algorithm to calculate the shortest distance between the two parts of pixel points, and using the median of the distance as the cell wall thickness; The loss function includes the cross-entropy loss function and the auxiliary loss function. The cross-entropy loss function is used to compare the difference between the output and the actual label to evaluate the prediction accuracy of the model at the pixel level. The auxiliary loss function is used to measure the similarity between the model's predicted output and the actual label to determine the degree of prediction error.

2. A method for calculating cell wall grayscale, hollowness, and thickness according to claim 1, characterized in that: The pretreatment includes: Read the ruler information, manually measure the ruler information, and calculate the actual length represented by the pixel in combination with the image pixel width for physical unit conversion of cell wall thickness parameters.

3. The method for calculating cell wall grayscale, hollowness, and thickness according to claim 1, wherein: The target segmentation model includes: The YOLO11 segmentation model is used. YOLO11 consists of a backbone network, a neck network, and a head network. The backbone network is used to extract image features, the neck network is used to further process and fuse image features, and the head network is used for target monitoring and classification. The target segmentation model is used to perform image segmentation, identify fungal cell walls, and extract cell wall contours.

4. A method for calculating cell wall grayscale, hollowness, and thickness according to claim 1, characterized in that: The data enhancement includes: Rotate, flip, and scale; Rotation, by rotating the image at random angles within a given parameter range to simulate shooting at different angles; Flip, by specifying the probability parameter of flip, the probability of flipping is controlled. The flips used include horizontal flipping and vertical flipping. Scaling: Given a parameter, the image is randomly scaled within the parameter range to simulate the performance of the target at different distances.

5. The method for calculating cell wall grayscale, hollowness, and thickness according to claim 1, wherein: The cell wall grayscale mean includes: The grayscale sum and total number of pixels in the cell wall area are extracted by masking, the grayscale value is calculated, and the average value is taken to reflect the overall brightness characteristics of the cell wall.

6. The method for calculating cell wall grayscale, hollowness, and thickness according to claim 1, wherein: The corner points include: Select target pixels near the edge of the cell image as a key reference for calculating the cell wall thickness. Determine the image edge and take the average value as the target edge. Traverse the pixel set and calculate the distance between each pixel and the target edge. Corner point selection includes the points closest to the left, right, top, and bottom edges of the image. By calculating the distance from the point to the edge, valid corner points are screened out.

7. The method for calculating cell wall grayscale, hollowness, and thickness according to claim 1, wherein: The algorithm optimization includes: A visual evaluation method based on CDF curve was used to verify the effectiveness and rationality of the algorithm by comparing and analyzing the parameter changes before and after the cell wall was decomposed by enzymes.

8. A cell wall grayscale, hollowness, and thickness calculation system, characterized by: Including image preprocessing module, target segmentation module, parameter calculation module; The image preprocessing module is used to label the input cell image and calculate the scale information to generate the cell wall position label; The target segmentation module is used to build and train the target segmentation algorithm model, optimize the loss function of image segmentation, and extract the cell wall contour; The parameter calculation module is used to calculate the grayscale mean, hollowing ratio and thickness of the cell wall based on the segmented cell image, and perform data enhancement and algorithm optimization; the hollowing ratio includes extracting the grayscale value of each pixel in the cell wall area through a mask, determining that the pixel is white, calculating the total number of white pixels and the total number of pixels, and calculating the hollowing ratio to characterize the internal integrity and degree of damage of the cell wall; the thickness includes segmenting the cell wall from the cell through a target segmentation model to obtain an edge pixel set, selecting a corner point, determining the effective corner point by calculating the distance between the corner point and the edge of the image, and using the KDTree algorithm to calculate the shortest distance between the two parts of the pixel points, and taking the median of the distance as the cell wall thickness.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of a method for calculating cell wall grayscale, hollowness, and thickness according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of a method for calculating cell wall grayscale, hollowness, and thickness according to any one of claims 1 to 7 are implemented.

Images