A multi-channel fluorescence image crosstalk adaptive correction method

Abstract

This invention discloses an adaptive correction method for crosstalk in multi-channel fluorescence images, relating to the field of fluorescence microscopy imaging technology. The method extracts images of the channels affected by crosstalk and the source channels of crosstalk from a multi-channel image sequence, determines candidate regions based on the pixel intensity distribution of the source channel image, iterates through crosstalk coefficients within a preset search range, and determines the crosstalk coefficients of the current frame with the goal of minimizing the standard deviation of the residual image. It then adaptively updates the crosstalk coefficients of multi-frame data using a time smoothing mechanism. Finally, it uses the updated crosstalk coefficients to complete crosstalk correction and performs negative value truncation on the corrected image. This invention eliminates the need for manual coefficient setting, exhibits high robustness, can adaptively cope with long-term drift and fluorescence attenuation changes, and can run in real-time on most microscope imaging systems.

Description

Technical Field

[0001] This invention relates to the field of fluorescence microscopy imaging technology, and more specifically to a method for adaptive correction of crosstalk in multi-channel fluorescence images. Background Technology

[0002] In fluorescence microscopy imaging, the overlap between the excitation and emission spectra of different fluorescent dyes leads to crosstalk between different channels. For example, in dual-channel imaging, the fluorescence signal of the C3 channel image may partially crosstalk into the C2 channel image, resulting in the C2 channel signal containing a crosstalk component from C3. Existing methods typically use a fixed crosstalk coefficient for simple linear subtraction; however, due to variations in spectral overlap between different samples, frames, and illumination conditions, the fixed coefficient method cannot effectively eliminate crosstalk. The results are as follows: in some frames, undersubtraction is insufficient, leaving crosstalk; in other frames, oversubtraction leads to an erroneous attenuation of the true C2 signal. Therefore, a method is needed that can adaptively determine the crosstalk coefficient based on the current image content, thereby stably removing crosstalk in long-term imaging or under different sample conditions. Summary of the Invention

[0003] In view of this, the present invention provides a multi-channel fluorescence image crosstalk adaptive correction method. To solve the above problems, the present invention adopts the following technical solution:

[0004] S1. Obtain a multi-channel fluorescence image sequence, wherein the multi-channel fluorescence image sequence includes multiple frames of images;

[0005] S2. Based on the multi-channel fluorescence image sequence, obtain the channel image affected by crosstalk and the channel image from which the crosstalk originates;

[0006] S3. Based on the pixel intensity distribution of the crosstalk source channel image, a candidate pixel intensity range is preset, and the regions corresponding to the candidate pixel intensity range in the crosstalk source channel image and the channel image affected by crosstalk are extracted as candidate regions.

[0007] S4. Based on the image statistical features of the candidate region, search within the preset crosstalk coefficient range to determine the crosstalk coefficient of the current frame;

[0008] S5. Combine the crosstalk coefficients of the current frame with the crosstalk coefficients of the previous frame, and generate the final crosstalk coefficients for correction through smoothing processing.

[0009] S6. Use the final crosstalk coefficient to perform crosstalk correction on the channel image affected by crosstalk to obtain the corrected image.

[0010] Preferably, in step S3, the preset candidate pixel intensity range includes:

[0011] The source image of the color-crossing channel is divided into multiple sub-images. The average intensity values ​​of the multiple sub-images are calculated and sorted. The image region coordinate range corresponding to the sub-image with the largest average intensity value is selected as the candidate pixel intensity range.

[0012] Preferably, in step S4, the image statistical features include the original signal of the channel image of the candidate region affected by crosstalk, the signal of the channel image from which the crosstalk originates in the candidate region, and the true signal of the channel image of the candidate region affected by crosstalk.

[0013] Furthermore, it also includes linear modeling based on the aforementioned image statistical features:

[0014]

[0015] in, The original signal of the channel image of the candidate region affected by crosstalk; The true signal of the channel image of the candidate region affected by crosstalk; The signal from the source channel image of the candidate region is cross-colored. These are the candidate color coefficient values.

[0016] Preferably, in step S4, determining the crosstalk coefficients of the current frame includes:

[0017] S42. Traverse the candidate color mixing coefficient values ​​using a fixed step size within the preset color mixing coefficient range;

[0018] S43. For each candidate cross-color coefficient value, calculate the statistical index of the residual image of the candidate region under the candidate cross-color coefficient value;

[0019] S44. Select the candidate crosstalk coefficient value that optimizes the statistical index as the crosstalk coefficient of the current frame.

[0020] Furthermore, the process of traversing the candidate color coefficient values ​​using a fixed step size involves at least two rounds, with the step size of the later round being smaller than that of the earlier round.

[0021] Furthermore, the statistical indicator is the standard deviation of the residual image, and the optimization is to minimize the standard deviation;

[0022] The formula for calculating the standard deviation is:

[0023]

[0024] The standard deviation of the residual image is represented by... This indicates the calculation of the standard deviation; The original signal of the channel image representing the candidate region affected by crosstalk; Represents the candidate color mixing coefficient; The signal representing the crosstalk source channel image of the candidate region;

[0025] Take the standard deviation of the residual image The coefficient corresponding to the minimum value As the color mixing factor for the current frame, that is:

[0026] in, This represents the color mixing factor for the current frame. Minimum parameters of the objective function .

[0027] Preferably, the process further includes negative value truncation of the corrected image.

[0028] Preferably, the multi-channel fluorescence image sequence includes dual-channel and above fluorescence image sequences. When it is multi-channel, for each channel affected by crosstalk, the corresponding crosstalk source channel is selected, and steps S1-S6 are repeated to complete the crosstalk correction for each channel.

[0029] The present invention also discloses a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of the multi-channel fluorescence image crosstalk adaptive correction method as described in any of the preceding claims.

[0030] As can be seen from the above technical solution, compared with the prior art, the beneficial effects of the present invention include:

[0031] 1. The optimal crosstalk coefficient can be automatically determined based on the principle of statistical self-consistency without the need for known standard samples;

[0032] 2. With the addition of a time smoothing mechanism, it can adaptively cope with long-term drift and fluorescence decay changes. Attached Figure Description

[0033] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0034] Figure 1 A flowchart of a multi-channel fluorescence image crosstalk adaptive correction method provided in an embodiment of the present invention;

[0035] Figure 2 This is a curve showing the variation of the standard deviation (std) of the residuals in the candidate region under different color mixing coefficients (k) provided in an embodiment of the present invention.

[0036] Figure 3 This document provides a table of standard deviation values ​​of candidate region residuals under different color mixing coefficients k, along with an explanation of the correction effect, for embodiments of the present invention. Detailed Implementation

[0037] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0038] In one embodiment, the present invention includes the following steps:

[0039] S1. Obtain a multi-channel fluorescence image sequence, which includes multiple frames of images;

[0040] S2. Based on the multi-channel fluorescence image sequence, obtain the channel image affected by crosstalk and the channel image of the crosstalk source;

[0041] S3. Based on the pixel intensity distribution of the source channel image of crosstalk, preset the candidate pixel intensity range, and extract the region corresponding to the candidate pixel intensity range in the source channel image of crosstalk and the channel image affected by crosstalk as the candidate region.

[0042] S4. Based on the image statistical features of the candidate region, search within the preset range of crosstalk coefficients to determine the crosstalk coefficient of the current frame.

[0043] S5. Combine the crosstalk coefficients of the current frame with the crosstalk coefficients of the previous frame, and generate the final crosstalk coefficients for correction through smoothing processing.

[0044] S6. Use the final crosstalk coefficient to perform crosstalk correction on the channel image affected by crosstalk to obtain the corrected image.

[0045] In one embodiment, the preset candidate pixel intensity range in step S3 includes:

[0046] The source channel image of the color smear is divided into multiple sub-images. The average intensity values ​​of the multiple sub-images are calculated and sorted. The coordinate range of the image region corresponding to the sub-image with the largest average intensity value is selected as the candidate pixel intensity range.

[0047] Specifically, a multi-channel image sequence acquired by a microscope is obtained. Taking dual-channel data as an example, the two channels are the channel image affected by crosstalk and the channel image from which crosstalk originates. Based on the pixel intensity distribution of the channel image from which crosstalk originates, regions with high signal-to-noise ratios and dominated by crosstalk signals are selected as candidate regions from the channel image affected by crosstalk. Specifically, the channel image from which crosstalk originates is divided into m*n sub-images, and the average intensity value of each sub-image is calculated. The sub-image with the largest average intensity value is selected as the candidate region.

[0048] In one embodiment, the image statistical features in step S4 include the original signal of the channel image of the candidate region affected by crosstalk, the signal of the channel image from which the crosstalk originates in the candidate region, and the true signal of the channel image of the candidate region affected by crosstalk.

[0049] Furthermore, a linear model is constructed based on image statistical features:

[0050]

[0051] in, The original signal of the channel image of the candidate region affected by color crosstalk. This represents the true signal of the channel image of the candidate region affected by color crosstalk. The signal from the source channel image of the candidate region is cross-colored. These are candidate color mixing coefficient values. When... When the value is close to the true crosstalk ratio, the spurious signal in channel C2 (the crosstalk portion from C3) is eliminated to the greatest extent, resulting in a corrected image. The overall intensity distribution tends to be uniform.

[0052] In one embodiment, determining the crosstalk coefficients of the current frame in step S4 includes:

[0053] S42. Traverse the candidate color mixing coefficient values ​​using a fixed step size within the preset range of color mixing coefficients;

[0054] S43. For each candidate cross-color coefficient value, calculate the statistical index of the residual image of the candidate region under the candidate cross-color coefficient value;

[0055] S44. Select the candidate crosstalk coefficient value that optimizes the statistical index as the crosstalk coefficient of the current frame.

[0056] Within the preset color bleeding coefficient range For example, within [0.1, 2.0], with a fixed step size. Traverse possible candidate color coefficient values Specifically, the preset range of color mixing coefficients. It can be determined based on the average intensity value of the source channel image and the color distortion source image, by default.

[0057]

[0058] The mean() function calculates the average intensity value of the candidate region. The original signal of the channel image of the candidate region affected by color crosstalk. For the signal of the source channel image of crosstalk,

[0059] Furthermore, the process of traversing the candidate color coefficient values ​​using a fixed step size takes at least two rounds, and the step size of the second round is smaller than that of the first round. That is, a second refinement search or fitting of the minimum point is performed near the minimum value of the first traversal.

[0060] Furthermore, the statistical indicator is the standard deviation of the residual image, and the optimization is to minimize the standard deviation.

[0061] The formula for calculating standard deviation is:

[0062]

[0063] The standard deviation of the residual image is represented by... This indicates the calculation of the standard deviation;

[0064] Take the standard deviation of the residual image The coefficient corresponding to the minimum value As the color mixing factor for the current frame, that is:

[0065] in, This represents the color mixing factor for the current frame. Minimum parameters of the objective function ;

[0066] The color mixing coefficient of the current frame is weighted and averaged with that of the previous frame:

[0067]

[0068] in This is the final color mixing coefficient; This is a smoothing coefficient, with a value range of [0.3, 0.7]. This represents the color mixing coefficient for the current frame; This is the color mixing coefficient of the previous frame;

[0069] In one embodiment, after color crosstalk correction, the corrected image is truncated with negative values.

[0070]

[0071] in, The true signal of the channel image of the candidate region affected by crosstalk; The original signal of the channel image of the candidate region affected by crosstalk; This is the final color mixing coefficient; This is the signal from the source channel image of the cross-color source.

[0072] like If a negative value is found, the corrected image is truncated by setting the negative value of the corrected image to 0;

[0073] In one embodiment, the multi-channel fluorescence image sequence includes dual-channel and above fluorescence image sequences. When it is multi-channel, for each channel affected by crosstalk, the corresponding crosstalk source channel is selected, and steps S1-S6 are repeated to complete the crosstalk correction for each channel.

[0074] In the specific implementation process, Figure 2 The curves show the variation of the standard deviation (std) of the residuals in the candidate region under different crosstalk coefficients (the horizontal axis represents the crosstalk coefficient, and the vertical axis represents the standard deviation of the residuals, std). The curves show that when the crosstalk coefficient is small, crosstalk is insufficiently eliminated, resulting in a large standard deviation of the residuals; when the crosstalk coefficient is large, crosstalk is excessively eliminated, also resulting in a large standard deviation of the residuals. At a certain optimal crosstalk coefficient value, the standard deviation of the residuals reaches its minimum, indicating that crosstalk is eliminated to the greatest extent and the correction effect is optimal. This curve visually verifies the effectiveness of determining the crosstalk coefficient by minimizing the standard deviation of the residuals.

[0075] Figure 3 Furthermore, a table of residual standard deviations corresponding to different crosstalk coefficients is provided, along with a schematic diagram of the correction effect. The table lists multiple crosstalk coefficient values ​​and their corresponding residual standard deviations, facilitating the rapid identification of the optimal crosstalk coefficient value. The schematic diagram compares the effects of uncorrected, undercorrected, overcorrected, and optimally corrected images, clearly demonstrating that the method of this invention can accurately remove crosstalk, preserve the true signal, and avoid errors caused by manually setting coefficients after adaptively determining the final crosstalk coefficients.

[0076] Based on the same inventive concept, embodiments of the present invention also provide a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of any of the aforementioned multi-channel fluorescence image crosstalk adaptive correction methods.

[0077] The above provides a detailed description of the multi-channel fluorescence image crosstalk adaptive correction method provided by the present invention. Specific examples are used in this embodiment to illustrate the principle and implementation of the present invention. The description of the above embodiments is only for the purpose of helping to understand the method and core idea of ​​the present invention. At the same time, for those skilled in the art, there will be changes in the specific implementation and application scope based on the idea of ​​the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

[0078] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined in these embodiments may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for adaptive color crosstalk correction in multi-channel fluorescence images, characterized in that, Includes the following steps: S1. Obtain a multi-channel fluorescence image sequence, wherein the multi-channel fluorescence image sequence includes multiple frames of images; S2. Based on the multi-channel fluorescence image sequence, obtain the channel image affected by crosstalk and the channel image from which the crosstalk originates; S3. Based on the pixel intensity distribution of the crosstalk source channel image, a candidate pixel intensity range is preset, and the regions corresponding to the candidate pixel intensity range in the crosstalk source channel image and the channel image affected by crosstalk are extracted as candidate regions. The preset candidate pixel intensity range includes: The color source channel image is divided into multiple sub-images, the average intensity values ​​of the multiple sub-images are calculated and sorted, and the image region coordinate range corresponding to the sub-image with the largest average intensity value is selected as the candidate pixel intensity range. S4. Based on the image statistical features of the candidate region, search within the preset crosstalk coefficient range to determine the crosstalk coefficient of the current frame; Determining the color mixing coefficient of the current frame includes: S42. Traverse the candidate color mixing coefficient values ​​using a fixed step size within the preset color mixing coefficient range; S43. For each candidate cross-color coefficient value, calculate the statistical index of the residual image of the candidate region under the candidate cross-color coefficient value; S44. Select the candidate crosstalk coefficient value that optimizes the statistical index as the crosstalk coefficient of the current frame. The statistical indicator is the standard deviation of the residual image, and the optimization is to minimize the standard deviation. The formula for calculating the standard deviation is: The standard deviation of the residual image is represented by... This indicates the calculation of the standard deviation; The original signal of the channel image representing the candidate region affected by crosstalk; Represents the candidate color mixing coefficient; The signal representing the crosstalk source channel image of the candidate region; Take the standard deviation of the residual image The coefficient corresponding to the minimum value As the color mixing factor for the current frame, that is: in, This represents the color mixing factor for the current frame. Minimum parameters of the objective function ; S5. Combine the crosstalk coefficients of the current frame with the crosstalk coefficients of the previous frame, and generate the final crosstalk coefficients for correction through smoothing processing. S6. Use the final crosstalk coefficient to perform crosstalk correction on the channel image affected by crosstalk to obtain the corrected image.

2. The adaptive correction method for color crosstalk in multi-channel fluorescence images according to claim 1, characterized in that, In step S4, the image statistical features include the original signal of the channel image of the candidate region affected by crosstalk, the signal of the channel image from which the crosstalk originates in the candidate region, and the true signal of the channel image of the candidate region affected by crosstalk.

3. The adaptive correction method for color crosstalk in multi-channel fluorescence images according to claim 2, characterized in that, It also includes linear modeling based on the aforementioned image statistical features: in, The original signal of the channel image of the candidate region affected by crosstalk; The true signal of the channel image of the candidate region affected by crosstalk; The signal from the source channel image of the candidate region is cross-colored. These are the candidate color coefficient values.

4. The adaptive correction method for color crosstalk in multi-channel fluorescence images according to claim 1, characterized in that, The process of traversing the candidate color coefficient values ​​using a fixed step size involves at least two rounds, with the step size of the later round being smaller than that of the earlier round.

5. The adaptive correction method for color crosstalk in multi-channel fluorescence images according to claim 1, characterized in that, Also includes: The corrected image is then truncated with negative values.

6. The adaptive correction method for color crosstalk in multi-channel fluorescence images according to claim 1, characterized in that, The multi-channel fluorescence image sequence includes dual-channel and above fluorescence image sequences. When it is multi-channel, for each channel affected by crosstalk, the corresponding crosstalk source channel is selected, and steps S1-S6 are repeated to complete the crosstalk correction for each channel.

7. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps of the multi-channel fluorescence image crosstalk adaptive correction method as described in any one of claims 1 to 6.

Images