An Image Processing Method for Fluorescent Probe Detection Based on Noise Analysis

Abstract

This invention relates to the field of image processing, and more particularly to an image processing method for fluorescence probe detection based on noise analysis. The method includes: acquiring a fluorescence microscopic grayscale image; obtaining a first optimization factor for the window corresponding to each pixel based on the local grayscale standard deviation within an initial window; obtaining a preliminary adaptive window side length based on the first optimization factor; obtaining the local contrast of the pixel based on the preliminary adaptive window; obtaining a second optimization factor for the window based on the first optimization factor and the local contrast; obtaining an adaptive local window side length based on the first optimization factor and the second optimization factor; obtaining a local adaptive contrast threshold for the pixel based on the first optimization factor, the second optimization factor, and the local contrast; and completing image segmentation processing based on the adaptive local window and the local adaptive contrast threshold, thereby improving the accuracy of image segmentation and enhancing the detection capability for weak signals.

Description

Technical Field

[0001] This invention relates to the field of image processing technology, and in particular to an image processing method for fluorescence probe detection based on noise analysis. Background Technology

[0002] Fluorescence microscopy plays a crucial role in modern biomedical research. By labeling specific cellular structures or biomolecules with fluorescent probes, researchers can observe the dynamic changes of various biological processes in living cells in real time. For example, labeling mitochondrial probes allows observation of mitochondrial fusion, division, and movement; labeling lysosomal probes allows study of intracellular degradation and transport processes. However, fluorescence microscopy images, especially those from living cells, often suffer from complex background noise, uneven signal intensity, and diverse organelle morphologies and sizes, posing challenges to subsequent image processing and quantitative analysis. Accurate extraction of regions of interest from fluorescence microscopy images requires image segmentation. Commonly used image segmentation methods include thresholding, edge detection, region growing, and cluster analysis. Among these, locally adaptive thresholding methods offer advantages in handling images with non-uniform lighting and complex backgrounds because they dynamically adjust the threshold based on local image features. The Bernsen algorithm is a classic locally adaptive thresholding algorithm that determines the threshold for each pixel by calculating the maximum and minimum gray values ​​within a local window surrounding that pixel. However, the traditional Bernson algorithm has some problems, such as sensitivity to noise and susceptibility to artifacts in noisy regions. In recent years, researchers have proposed many improved Bernson algorithms to improve their robustness to noise. Some methods introduce contrast constraints, using the threshold calculated by the Bernson algorithm only when the local contrast is greater than a certain preset value; otherwise, a global threshold or other methods are used to calculate the threshold. Other methods combine neighborhood information, considering not only the local window around the current pixel but also the threshold information of its neighboring pixels when calculating the local threshold. These improved methods have improved the segmentation accuracy to some extent, but a core problem remains: they typically use fixed-size local windows and pre-set contrast thresholds, making it difficult to adapt to the diverse organelle sizes and non-uniform signal intensity prevalent in fluorescence microscopy images. Existing improved Bernson algorithms, with their fixed-size local windows and fixed contrast thresholds, struggle to adapt to the characteristics of organelle size diversity and non-uniform signal intensity in fluorescence microscopy images, leading to inaccurate segmentation in noisy, low-contrast regions, and a tendency to lose weak signals or produce artifacts.

[0003] Specifically, fixed-size local windows struggle to accommodate organelles of varying sizes. For larger organelles, smaller windows fail to fully encompass them, leading to incomplete segmentation. Conversely, for smaller organelles, larger windows include excessive background noise, resulting in inaccurate threshold calculations. Furthermore, background noise levels and signal intensities often differ significantly across regions in fluorescence microscopy images. Fixed-size windows and fixed contrast thresholds cannot simultaneously adapt to these variations, resulting in poor segmentation in certain areas. Particularly in applications such as organelle tracking, the loss of weak signals can disrupt or erroneously alter the tracking trajectory, severely impacting subsequent quantitative analysis.

[0004] Therefore, there is an urgent need for a processing method that can adaptively adjust the size of the local window and the contrast threshold based on the local features of fluorescence microscopy images, thereby improving the accuracy of image segmentation and enhancing the ability to detect weak signals. Summary of the Invention

[0005] In view of this, embodiments of the present invention provide a fluorescent probe detection image processing method based on noise analysis to solve the problem that fixed-size windows and fixed contrast thresholds cannot simultaneously adapt to these changes, resulting in poor segmentation results in certain regions.

[0006] This invention provides an image processing method for fluorescent probe detection based on noise analysis. The method includes the following steps:

[0007] The process involves acquiring fluorescence microscopy images of the cells to be tested; preprocessing and converting the acquired fluorescence microscopy images to obtain fluorescence microscopy grayscale images; analyzing the local grayscale changes of pixels within an initial window in the fluorescence microscopy grayscale image to obtain a first optimization factor for the pixel window; adjusting the initial window of the pixel using the first optimization factor to obtain the preliminary adaptive window side length of the pixel in the fluorescence microscopy grayscale image; performing contrast analysis on the preliminary adaptive window of the pixel in the fluorescence microscopy grayscale image to obtain the local contrast of the pixel; analyzing the local contrast of the pixel to obtain a second optimization factor for the pixel window in the fluorescence microscopy grayscale image; evaluating the window of the pixel in the fluorescence microscopy grayscale image using the first and second optimization factors to obtain the adaptive local window side length of the pixel; analyzing the window content of the adaptive local window of the pixel to obtain the local adaptive contrast threshold of the pixel; and segmenting the fluorescence microscopy grayscale image using the adaptive local window of the pixel and the local adaptive contrast threshold to obtain a binary segmented image corresponding to the fluorescence microscopy grayscale image.

[0008] Preferably, the method for obtaining the first optimization factor of the pixel window by analyzing the local gray-level changes of pixels within an initial window in a fluorescence microscopy gray-scale image includes the following specific steps:

[0009] Obtain the set window adjustment factor, window optimization regularization factor, and initial window size. Divide the pixels in the fluorescence microscopy image into local windows using the initial window size to obtain the first local window corresponding to the pixel in the fluorescence microscopy image. Divide the window adjustment factor by the sum of the standard deviation of the gray value of the pixel in the first local window corresponding to the pixel in the fluorescence microscopy image and the window optimization regularization factor as the first optimization factor of the window corresponding to the pixel.

[0010] Preferably, the initial adaptive window side length of the pixels in the fluorescence microscopy grayscale image is obtained by adjusting the initial window of the pixels according to the first optimization factor of the window, and the specific steps include:

[0011] Obtain the set window limit constant, multiply the window limit constant by the first window optimization factor and round down to get the result as the first window side length, and multiply the constant 2 by the first window side length and add it to the constant 3 to get the result as the initial adaptive window side length corresponding to the pixel.

[0012] Preferably, the method for obtaining the local contrast of pixels by performing contrast analysis on a preliminary adaptive window of pixels in a fluorescence microscopy grayscale image includes the following specific steps:

[0013] Obtain the initial adaptive window side length corresponding to the pixel. Pixels in the initial adaptive window whose grayscale value is greater than the average grayscale value of all pixels within the window are designated as the first set of pixels corresponding to the pixel, and the average grayscale value of the pixels in the first set of pixels is designated as the first average grayscale value corresponding to the pixel. Pixels in the initial adaptive window whose grayscale value is less than the average grayscale value of all pixels within the window are designated as the second set of pixels corresponding to the pixel, and the average grayscale value of the pixels in the second set of pixels is designated as the second average grayscale value corresponding to the pixel. The result of subtracting the first average grayscale value from the second average grayscale value corresponding to the pixel is used as the local contrast of the pixel.

[0014] Preferably, the method for obtaining the second window optimization factor for pixels in the fluorescence microscopy grayscale image by analyzing the local contrast of pixels includes the following specific steps:

[0015] Obtain the first optimization factor of the window corresponding to the pixel and the local contrast corresponding to the pixel, and use the calculation result of dividing the local contrast corresponding to the pixel by the first optimization factor of the window corresponding to the pixel as the second optimization factor of the window corresponding to the pixel.

[0016] Preferably, the method for evaluating the pixels in the fluorescence microscopy grayscale image using the first and second window optimization factors to obtain the adaptive local window side length for each pixel includes the following steps:

[0017] Obtain the set window limit constant, the first optimization factor of the window corresponding to the pixel, and the second optimization factor of the window corresponding to the pixel; multiply the window limit constant, the first optimization factor of the window corresponding to the pixel, and the second optimization factor of the window corresponding to the pixel and round down to obtain the result as the second window side length; multiply constant 2 by the second window side length and add it to constant 3 to obtain the result as the adaptive local window side length corresponding to the pixel.

[0018] Preferably, obtaining the local adaptive contrast threshold of the pixel based on the first window optimization factor corresponding to the pixel, the second window optimization factor corresponding to the pixel, and the local contrast corresponding to the pixel includes:

[0019] Obtain the first optimization factor of the window corresponding to the pixel, the second optimization factor of the window corresponding to the pixel, and the local contrast corresponding to the pixel; divide the first optimization factor of the window corresponding to the pixel by the second optimization factor of the window corresponding to the pixel as the contrast threshold evaluation factor, and add a constant to the contrast threshold evaluation factor and multiply by the local contrast corresponding to the pixel as the local adaptive contrast threshold of the pixel.

[0020] Preferably, the step of obtaining a binary segmented image corresponding to the fluorescence microscopy grayscale image by performing fluorescence microscopy grayscale image segmentation based on the adaptive local window side length corresponding to the pixel and the local adaptive contrast threshold of the pixel includes:

[0021] The adaptive local window side length and the local adaptive contrast threshold of the pixel are obtained. The local adaptive window of the pixel in the fluorescence microscopy grayscale image is determined according to the adaptive local window side length. The average gray value of all pixels in the local adaptive window and the local adaptive contrast threshold are added to the pixel as the segmentation threshold. The gray value of the pixel is compared with the segmentation threshold to obtain the binary segmentation label of the pixel. The binary segmentation of the fluorescence microscopy image is completed, and the binary segmented image corresponding to the fluorescence microscopy grayscale image is obtained.

[0022] Preferably, the step of comparing the grayscale value of the pixel with the segmentation threshold of the pixel to obtain the binary segmentation label of the pixel includes:

[0023] Obtain the segmentation threshold of the pixel. If the gray value of the pixel is greater than or equal to the segmentation threshold, the pixel is marked as a target pixel. If the gray value of the pixel is less than the segmentation threshold, the pixel is marked as a background pixel.

[0024] The beneficial effects of the embodiments of the present invention compared with the prior art are as follows:

[0025] This invention involves acquiring fluorescence microscopic images of cells to be detected; preprocessing and converting the acquired fluorescence microscopic images to grayscale to obtain fluorescence microscopic grayscale images; analyzing the local grayscale changes of pixels within an initial window in the fluorescence microscopic grayscale image to obtain a first optimization factor for the pixel window; adjusting the initial window of the pixel using the first optimization factor to obtain a preliminary adaptive window side length for the pixel in the fluorescence microscopic grayscale image; performing contrast analysis on the preliminary adaptive window of the pixel in the fluorescence microscopic grayscale image to obtain the local contrast of the pixel; analyzing the local contrast of the pixel to obtain a second optimization factor for the pixel window in the fluorescence microscopic grayscale image; evaluating the window of the pixel in the fluorescence microscopic grayscale image using the first and second optimization factors to obtain an adaptive local window side length for the pixel; analyzing the window content of the adaptive local window of the pixel to obtain a local adaptive contrast threshold for the pixel; and segmenting the fluorescence microscopic grayscale image using the adaptive local window of the pixel and the local adaptive contrast threshold to obtain a binary segmented image corresponding to the fluorescence microscopic grayscale image. Specifically, a first optimization factor for the window is obtained by using the standard deviation of the gray values ​​of all pixels within an initial window in the fluorescence microscopy image. A preliminary adaptive window is then obtained based on the first optimization factor. A second optimization factor for the window is evaluated using the preliminary adaptive window. Subsequently, a local adaptive window for Bernstein binary segmentation of the fluorescence microscopy image is obtained using the first and second optimization factors. Finally, an adaptive image segmentation threshold is obtained based on the first and second optimization factors and the local contrast within the local adaptive window, thereby improving the accuracy of image segmentation and enhancing the detection capability of weak signals. Attached Figure Description

[0026] The accompanying drawings, which constitute part of the present invention, are intended to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are intended to explain the present invention and do not constitute an undue limitation of the present invention. In the accompanying drawings:

[0027] Figure 1 This is a flowchart of an image processing method for fluorescence probe detection based on noise analysis, as described in an embodiment of the present invention. Detailed Implementation

[0028] Embodiments of this disclosure are described in detail below, examples of which are illustrated in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this disclosure, and should not be construed as limiting this disclosure.

[0029] It should be noted that the terms "first," "second," etc., used in the specification and accompanying drawings of this disclosure are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this toothpaste application can be interchanged where appropriate so that the embodiments of this disclosure described herein can be implemented in orders other than those illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this disclosure. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this disclosure.

[0030] To illustrate the technical solution of the present invention, specific embodiments are described below.

[0031] The specific scenario addressed by this invention is image segmentation of microscopic images detected by fluorescent probes.

[0032] See Figure 1 This is a flowchart of an image processing method for fluorescent probe detection based on noise analysis, provided in Embodiment 1 of the present invention. Figure 1 As shown, an image processing method for fluorescent probe detection based on noise analysis may include:

[0033] Step S101: Acquire fluorescence microscopic images of cells, preprocess the acquired cell fluorescence microscopic images and convert them into grayscale images to obtain fluorescence microscopic grayscale images.

[0034] The cell fluorescence microscopy image is preprocessed using the top-hat transformation method to obtain a background-corrected cell fluorescence microscopy image. The background-corrected cell fluorescence microscopy image is then subjected to grayscale transformation to obtain a cell fluorescence microscopy grayscale image.

[0035] Step S102: Obtain the first optimization factor of the window corresponding to the pixel in the initial window based on the local grayscale standard deviation of the pixel in the fluorescence microscopy grayscale image; obtain the preliminary adaptive window side length corresponding to the pixel based on the first optimization factor of the window corresponding to the pixel.

[0036] In fluorescence microscopy image processing of cells, especially for dynamic tracking of organelles, accurate image segmentation is fundamental for subsequent quantitative analysis. However, organelles exhibit significant diversity in size and morphology, and background noise levels vary considerably across different regions of the image. Traditional image processing methods using a fixed window size based on Bernstein's image segmentation approach struggle to address both of these challenges. In noisy regions, a larger window can better estimate the background but sacrifices image detail, potentially misclassifying weak organelle signals as background. Conversely, a smaller window near organelle edges or smaller organelle structures better preserves edge details but is more susceptible to noise interference, leading to inaccurate segmentation results. Therefore, after acquiring the fluorescence microscopy grayscale image, it is necessary to obtain the first optimization factor for the window corresponding to each pixel within the initial window based on the local grayscale standard deviation of the pixel in the initial window. This factor includes:

[0037] Obtain the set window adjustment factor, window optimization regularization factor, and initial window size. Divide the pixels in the fluorescence microscopy image into local windows using the initial window size to obtain the first local window corresponding to the pixel in the fluorescence microscopy image. Divide the window adjustment factor by the sum of the standard deviation of the gray value of the pixel in the first local window corresponding to the pixel in the fluorescence microscopy image and the window optimization regularization factor as the first optimization factor of the window corresponding to the pixel.

[0038] In one embodiment, it is assumed that the coordinates in the fluorescence microscopic grayscale image are... The standard deviation of grayscale values ​​in the initial window of the pixels is Then the expression for calculating the first optimization factor of the window corresponding to the pixel is:

[0039]

[0040] in, Indicates the coordinates in the image are The first optimization factor for the window of pixels; Indicates the window adjustment factor; This represents the window optimization regularization factor.

[0041] It should be noted that this step aims to initially address the problem of fixed local window sizes. The design idea for the first optimization factor of the pixel window is based on the analysis of the local gray-level statistical characteristics of the image, especially the local gray-level standard deviation. This is because of the local grayscale standard deviation. This can serve as an important indicator for measuring the noise level in a local area. Specifically, for each pixel in the image, the standard deviation of the gray values ​​within an initial window surrounding it is first calculated. In this embodiment, the initial window is set to 5x5. A larger value usually means that the grayscale changes in the area are drastic, there is strong noise, or it contains edge information of organelles; A smaller value usually means that the grayscale change in the area is gradual, mainly containing background signals or relatively uniform areas inside organelles.

[0042] The solution based on this understanding is to use the local grayscale standard deviation around each pixel. To dynamically adjust the window size. When the window size is large, appropriately reduce the window size to avoid introducing too much noise while better preserving the detailed information of organelle edges; when When the background is small, appropriately increasing the window size yields a more stable background estimate, thus enabling more accurate threshold calculation. To control the influence of the grayscale standard deviation on the window size, an adjustment factor needs to be introduced. . Its function is to adjust the sensitivity of the noise level to window resizing, similar to a "magnification factor". When Setting a large grayscale standard deviation can lead to a significant adjustment in window size, even with small changes. The specific value needs to be adjusted according to the noise level of different datasets.

[0043] The above method enables preliminary adaptive window size adjustment based on local grayscale standard deviation. Factors and They are inversely proportional. The window size is initially adjusted based on the local noise level, allowing the algorithm to initially adapt to the diversity of organelle sizes and the spatial heterogeneity of noise levels.

[0044] After obtaining the first window optimization factor for each pixel in the fluorescence microscopy grayscale image, it is necessary to obtain the preliminary adaptive window side length corresponding to the pixel based on the first window optimization factor, including:

[0045] Obtain the set window limit constant, multiply the window limit constant by the first window optimization factor and round down to get the result as the first window side length, and multiply the constant 2 by the first window side length and add it to the constant 3 to get the result as the initial adaptive window side length corresponding to the pixel.

[0046] In one implementation, it is assumed that the window limit constant is... The initial adaptive window side length calculation expression for each pixel is:

[0047]

[0048] in, The coordinates in the fluorescence microscopy grayscale image are The initial adaptive window side length for each pixel; Indicates the window limit constant; The coordinates in the fluorescence microscopy grayscale image are The first optimization factor for the window of pixels; This represents the floor function.

[0049] It should be noted that in this embodiment, the window limit constant is set to... This ensures that the minimum window size is... and through To ensure that the window side length is odd, a window limit constant is introduced. It serves to control the overall size of the window, because The value is greater than ,so The minimum value is , then when The result of this part is In order to minimize the window size ,need .

[0050] Step S103: Obtain the local contrast of the pixel according to the preliminary adaptive window corresponding to the pixel; obtain the second optimization factor of the window corresponding to the pixel according to the first optimization factor of the window corresponding to the pixel and the local contrast of the pixel; obtain the side length of the adaptive local window corresponding to the pixel according to the first optimization factor of the window corresponding to the pixel and the second optimization factor of the window corresponding to the pixel.

[0051] In fluorescence microscopy images, the signal intensity of organelles often varies considerably due to factors such as the concentration of fluorescent dyes used to label organelles, the spatial distribution of organelles, and the characteristics of the optical imaging system. There are both bright areas with strong signals and dark areas with weak signals. If only the grayscale standard deviation is considered... Adjusting the window size can lead to situations where the window is too small in low-contrast areas, making it difficult to accurately estimate the grayscale level of the local background, thus affecting the accuracy of subsequent threshold calculations. This is especially problematic in applications like organelle tracking, where it's desirable to preserve as many weak organelle signals as possible. These weak signals often correspond to the initiation, termination, or regions of interaction with other organelles, and have significant biological implications.

[0052] Therefore, after obtaining the initial adaptive window side length for each pixel, further optimization is needed. Specifically, firstly, the local contrast corresponding to the pixel is obtained based on the initial adaptive window, including:

[0053] Obtain the initial adaptive window side length corresponding to the pixel. Pixels in the initial adaptive window whose grayscale value is greater than the average grayscale value of all pixels within the window are designated as the first set of pixels corresponding to the pixel, and the average grayscale value of the pixels in the first set of pixels is designated as the first average grayscale value corresponding to the pixel. Pixels in the initial adaptive window whose grayscale value is less than the average grayscale value of all pixels within the window are designated as the second set of pixels corresponding to the pixel, and the average grayscale value of the pixels in the second set of pixels is designated as the second average grayscale value corresponding to the pixel. The result of subtracting the first average grayscale value from the second average grayscale value corresponding to the pixel is used as the local contrast of the pixel.

[0054] In one embodiment, it is assumed that the coordinates in the fluorescence microscopy grayscale image are... The coordinates of the pixels in the initial adaptive window are The grayscale value of the pixel Then the expression for calculating the local contrast of the corresponding pixel is:

[0055]

[0056] in, This represents the local contrast of a pixel. The coordinates in the fluorescence microscopy grayscale image are The coordinates of the pixels in the initial adaptive window are The grayscale value of the pixel; An initial adaptive window representing pixels; Indicates in The average grayscale value of all pixels within the window; The coordinates in the fluorescence microscopy grayscale image are The number of pixels in the initial adaptive window that are greater than or equal to the mean gray value; The coordinates in the fluorescence microscopy grayscale image are The number of pixels smaller than the average grayscale value in the initial adaptive window of the pixels.

[0057] After obtaining the local contrast corresponding to the pixel, the second optimization factor of the window corresponding to the pixel is obtained based on the first optimization factor of the window corresponding to the pixel and the local contrast corresponding to the pixel, including:

[0058] Obtain the first optimization factor of the window corresponding to the pixel and the local contrast corresponding to the pixel, and use the calculation result of dividing the local contrast corresponding to the pixel by the first optimization factor of the window corresponding to the pixel as the second optimization factor of the window corresponding to the pixel.

[0059] In one embodiment, the expression for calculating the second optimization factor of the window corresponding to the pixel is:

[0060]

[0061] in, This represents the second optimization factor of the window corresponding to the pixel. The coordinates in the fluorescence microscopy grayscale image are The local contrast corresponding to each pixel; The coordinates in the fluorescence microscopy grayscale image are The first optimization factor for the window of pixels.

[0062] It should be noted that, and The calculation avoids the noise sensitivity problem caused by directly using the maximum and minimum grayscale values ​​within the window to calculate contrast. Local contrast. Reflected in The difference in grayscale between bright and dark areas within the window. Optimization factor. pass and The ratio is used to combine the local contrast information with the noise level information obtained in the previous steps. Specifically, in areas with high contrast or high noise, A larger value will lead to a further increase in the final window size, thus resulting in a better estimation of the background; while in areas with low contrast and low noise, The value will be smaller, thus avoiding an excessively large window size and helping to preserve weak signals.

[0063] After obtaining the second window optimization factor for a pixel, the adaptive local window side length corresponding to the pixel can be obtained based on the first window optimization factor and the second window optimization factor corresponding to the pixel, including:

[0064] Obtain the set window limit constant, the first optimization factor of the window corresponding to the pixel, and the second optimization factor of the window corresponding to the pixel; multiply the window limit constant, the first optimization factor of the window corresponding to the pixel, and the second optimization factor of the window corresponding to the pixel and round down to obtain the result as the second window side length; multiply constant 2 by the second window side length and add it to constant 3 to obtain the result as the adaptive local window side length corresponding to the pixel.

[0065] In one embodiment, the expression for calculating the adaptive local window side length corresponding to a pixel is:

[0066]

[0067] in, The coordinates in the fluorescence microscopy grayscale image are The adaptive local window side length for each pixel; Indicates the window limit constant; The coordinates in the fluorescence microscopy grayscale image are The first optimization factor for the window of pixels; The coordinates in the fluorescence microscopy grayscale image are The second optimization factor for the window of pixels.

[0068] Step S104: Obtain the local adaptive contrast threshold of the pixel based on the first optimization factor of the window corresponding to the pixel, the second optimization factor of the window corresponding to the pixel, and the local contrast corresponding to the pixel; perform fluorescence microscopy grayscale image segmentation based on the adaptive local window corresponding to the pixel and the local adaptive contrast threshold of the pixel to obtain the binary segmented image corresponding to the fluorescence microscopy grayscale image.

[0069] After obtaining the adaptive local window side length corresponding to the pixel, the local adaptive contrast threshold of the pixel can be obtained based on the first window optimization factor, the second window optimization factor, and the local contrast of the pixel, including:

[0070] Obtain the first optimization factor of the window corresponding to the pixel, the second optimization factor of the window corresponding to the pixel, and the local contrast corresponding to the pixel; divide the first optimization factor of the window corresponding to the pixel by the second optimization factor of the window corresponding to the pixel as the contrast threshold evaluation factor, and add a constant to the contrast threshold evaluation factor and multiply by the local contrast corresponding to the pixel as the local adaptive contrast threshold of the pixel.

[0071] In one embodiment, the expression for calculating the local adaptive contrast threshold of a pixel is:

[0072]

[0073] in, The coordinates in the fluorescence microscopy grayscale image are Local adaptive contrast threshold for pixels; The coordinates in the fluorescence microscopy grayscale image are Local contrast of pixels; The coordinates in the fluorescence microscopy grayscale image are The first optimization factor for the window of pixels; The coordinates in the fluorescence microscopy grayscale image are The second optimization factor for the window of pixels.

[0074] After obtaining the local adaptive contrast threshold and the adaptive local window of a pixel, fluorescence microscopy grayscale image segmentation can be performed based on the adaptive local window and the local adaptive contrast threshold of the pixel to obtain a binary segmented image corresponding to the fluorescence microscopy grayscale image, including:

[0075] The adaptive local window side length and the local adaptive contrast threshold of the pixel are obtained. The local adaptive window of the pixel in the fluorescence microscopy grayscale image is determined according to the adaptive local window side length. The average gray value of all pixels in the local adaptive window and the local adaptive contrast threshold are added to the pixel as the segmentation threshold. The gray value of the pixel is compared with the segmentation threshold to obtain the binary segmentation label of the pixel. The binary segmentation of the fluorescence microscopy image is completed, and the binary segmented image corresponding to the fluorescence microscopy grayscale image is obtained.

[0076] The process of comparing the grayscale value of a pixel with a segmentation threshold to obtain a binary segmentation label for the pixel includes:

[0077] A segmentation threshold is obtained for each pixel. If the grayscale value of the pixel is greater than or equal to the segmentation threshold, the pixel is marked as a target pixel; if the grayscale value of the pixel is less than the segmentation threshold, the pixel is marked as a background pixel. This completes the binary segmentation processing of the fluorescent probe-detected image.

[0078] In summary, the embodiments of the present invention acquire fluorescence microscopic images of cells to be detected; preprocess and convert the acquired fluorescence microscopic images to grayscale to obtain fluorescence microscopic grayscale images; analyze the local grayscale changes of pixels within an initial window in the fluorescence microscopic grayscale image to obtain a first optimization factor for the pixel window; adjust the initial window of the pixel using the first optimization factor to obtain the preliminary adaptive window side length of the pixel in the fluorescence microscopic grayscale image; perform contrast analysis on the preliminary adaptive window of the pixel in the fluorescence microscopic grayscale image to obtain the local contrast of the pixel; analyze the local contrast of the pixel to obtain a second optimization factor for the pixel window in the fluorescence microscopic grayscale image; evaluate the window of the pixel in the fluorescence microscopic grayscale image using the first and second optimization factors to obtain the adaptive local window side length of the pixel; analyze the window content of the adaptive local window of the pixel to obtain the local adaptive contrast threshold of the pixel; and segment the fluorescence microscopic grayscale image using the adaptive local window of the pixel and the local adaptive contrast threshold to obtain a binary segmented image corresponding to the fluorescence microscopic grayscale image. Specifically, a first optimization factor for the window is obtained by using the standard deviation of the gray values ​​of all pixels within an initial window in the fluorescence microscopy image. A preliminary adaptive window is then obtained based on the first optimization factor. A second optimization factor for the window is evaluated using the preliminary adaptive window. Subsequently, a local adaptive window for Bernstein binary segmentation of the fluorescence microscopy image is obtained using the first and second optimization factors. Finally, an adaptive image segmentation threshold is obtained based on the first and second optimization factors and the local contrast within the local adaptive window, thereby improving the accuracy of image segmentation and enhancing the detection capability of weak signals.

[0079] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A fluorescence probe detection image processing method based on noise analysis, comprising: Acquire fluorescence microscopic images of the cells to be tested; The acquired cell fluorescence microscopy images are preprocessed and converted to grayscale to obtain fluorescence microscopy grayscale images. The method is characterized by: analyzing the local grayscale changes of pixels within an initial window in the fluorescence microscopy grayscale image to obtain a first window optimization factor for each pixel; adjusting the initial window of each pixel using the first window optimization factor to obtain a preliminary adaptive window side length for each pixel in the fluorescence microscopy grayscale image; performing contrast analysis on the preliminary adaptive window of each pixel in the fluorescence microscopy grayscale image to obtain the local contrast of each pixel; analyzing the local contrast of each pixel to obtain a second window optimization factor for each pixel in the fluorescence microscopy grayscale image; evaluating the window of each pixel in the fluorescence microscopy grayscale image using the first and second window optimization factors to obtain an adaptive local window side length for each pixel; analyzing the window content of the adaptive local window of each pixel to obtain a local adaptive contrast threshold for each pixel; and segmenting the fluorescence microscopy grayscale image using the adaptive local window of each pixel and the local adaptive contrast threshold to obtain a corresponding binary segmented image of the fluorescence microscopy grayscale image.

2. The image processing method for fluorescence probe detection based on noise analysis according to claim 1, characterized in that, According to the above, by performing local gray-level change analysis on pixels within an initial window in a fluorescence microscopy grayscale image, the first optimization factor of the pixel window is obtained. The specific steps include: Obtain the set window adjustment factor, window optimization regularization factor, and initial window size. Divide the pixels in the fluorescence microscopy image into local windows using the initial window size to obtain the first local window corresponding to the pixel in the fluorescence microscopy image. Divide the window adjustment factor by the sum of the standard deviation of the gray value of the pixel in the first local window corresponding to the pixel in the fluorescence microscopy image and the window optimization regularization factor as the first optimization factor of the window corresponding to the pixel.

3. The image processing method for fluorescence probe detection based on noise analysis according to claim 1, characterized in that, The initial adaptive window side length of the pixels in the fluorescence microscopy grayscale image is obtained by adjusting the initial window of the pixels using the first optimization factor of the window. The specific steps include: Obtain the set window limit constant, multiply the window limit constant by the first window optimization factor and round down to get the result as the first window side length, and multiply the constant 2 by the first window side length and add it to the constant 3 to get the result as the initial adaptive window side length corresponding to the pixel.

4. The image processing method for fluorescence probe detection based on noise analysis according to claim 1, characterized in that, The specific steps for obtaining the local contrast of pixels by performing contrast analysis on a preliminary adaptive window of pixels in a fluorescence microscopy grayscale image include: Obtain the initial adaptive window side length corresponding to the pixel. Pixels in the initial adaptive window whose grayscale value is greater than the average grayscale value of all pixels within the window are designated as the first set of pixels corresponding to the pixel, and the average grayscale value of the pixels in the first set of pixels is designated as the first average grayscale value corresponding to the pixel. Pixels in the initial adaptive window whose grayscale value is less than the average grayscale value of all pixels within the window are designated as the second set of pixels corresponding to the pixel, and the average grayscale value of the pixels in the second set of pixels is designated as the second average grayscale value corresponding to the pixel. The result of subtracting the first average grayscale value from the second average grayscale value corresponding to the pixel is used as the local contrast of the pixel.

5. The image processing method for fluorescence probe detection based on noise analysis according to claim 1, characterized in that, The second optimization factor for the window of pixels in the fluorescence microscopy grayscale image is obtained by analyzing the local contrast of the pixels. The specific steps include: Obtain the first optimization factor of the window corresponding to the pixel and the local contrast corresponding to the pixel, and use the calculation result of dividing the local contrast corresponding to the pixel by the first optimization factor of the window corresponding to the pixel as the second optimization factor of the window corresponding to the pixel.

6. The image processing method for fluorescence probe detection based on noise analysis according to claim 1, characterized in that, The steps for evaluating the pixel values ​​in the fluorescence microscopy grayscale image using the first and second window optimization factors to obtain the adaptive local window side length for each pixel include: Obtain the set window limit constant, the first optimization factor of the window corresponding to the pixel, and the second optimization factor of the window corresponding to the pixel; multiply the window limit constant, the first optimization factor of the window corresponding to the pixel, and the second optimization factor of the window corresponding to the pixel and round down to obtain the result as the second window side length; multiply constant 2 by the second window side length and add it to constant 3 to obtain the result as the adaptive local window side length corresponding to the pixel.

7. The image processing method for fluorescence probe detection based on noise analysis according to claim 1, characterized in that, The specific steps for obtaining the local adaptive contrast threshold of a pixel by performing window content analysis on the adaptive local window of the pixel include: Obtain the first optimization factor of the window corresponding to the pixel, the second optimization factor of the window corresponding to the pixel, and the local contrast corresponding to the pixel; divide the first optimization factor of the window corresponding to the pixel by the second optimization factor of the window corresponding to the pixel as the contrast threshold evaluation factor, and add a constant to the contrast threshold evaluation factor and multiply by the local contrast corresponding to the pixel as the local adaptive contrast threshold of the pixel.

8. The image processing method for fluorescence probe detection based on noise analysis according to claim 1, characterized in that, The fluorescence microscopy grayscale image is segmented using an adaptive local window and a local adaptive contrast threshold to obtain a binary segmented image corresponding to the fluorescence microscopy grayscale image. The specific steps include: The adaptive local window side length and the local adaptive contrast threshold of the pixel are obtained. The local adaptive window of the pixel in the fluorescence microscopy grayscale image is determined according to the adaptive local window side length. The average gray value of all pixels in the local adaptive window and the local adaptive contrast threshold are added to the pixel as the segmentation threshold. The gray value of the pixel is compared with the segmentation threshold to obtain the binary segmentation label of the pixel. The binary segmentation of the fluorescence microscopy image is completed, and the binary segmented image corresponding to the fluorescence microscopy grayscale image is obtained.

9. The image processing method for fluorescence probe detection based on noise analysis according to claim 8, characterized in that, The specific steps for obtaining the binary segmentation label of a pixel by comparing its grayscale value with a segmentation threshold include: Obtain the segmentation threshold of the pixel. If the gray value of the pixel is greater than or equal to the segmentation threshold, the pixel is marked as a target pixel. If the gray value of the pixel is less than the segmentation threshold, the pixel is marked as a background pixel.

Images