CN-121746396-B - Fluorescence immunohistochemical automatic exposure value-taking method and system

CN121746396BCN 121746396 BCN121746396 BCN 121746396BCN-121746396-B

Abstract

The invention relates to the technical field of image data processing, in particular to a fluorescence immunohistochemical automatic exposure value-taking method and system. The method comprises the steps of controlling an excitation light source to continuously collect with preset micro-dose parameters, obtaining two frames of probe frames, denoising by utilizing a pre-stored dark field image to obtain a net signal image matrix, extracting gradient characteristics of the first net signal matrix, combining gray values to construct a form and light intensity combined confidence coefficient matrix, representing the probability that each pixel point is an effective biological tissue signal, carrying out weighted statistics on the net signal matrix by utilizing the confidence coefficient matrix to calculate fluorescence attenuation ratio, and reversely pushing theoretical maximum fluorescence intensity of a sample before photobleaching, and calculating optimal exposure time according to the theoretical intensity, a target gray value and probe time. The scheme of the invention can realize zero-loss protection of pathological samples, accurately extract effective signals from low signal-to-noise ratio background and improve analysis accuracy.

Inventors

Bu Lingbin
ZHANG RUI
HU ZICHAO
Su Jingun

Assignees

阔然生物医药科技(上海)有限公司
江苏阔然生物医药科技有限公司

Dates

Publication Date: 20260512
Application Date: 20260228

Claims (10)

1. The fluorescence immunohistochemical automatic exposure and value-taking method is characterized by comprising the following steps: the method comprises the steps of controlling an excitation light source to continuously acquire a current view field with preset micro-dose parameters, acquiring a first probe frame and a second probe frame, and processing the first probe frame and the second probe frame by utilizing pre-stored dark field images to acquire a denoised first clean signal image matrix and a denoised second clean signal image matrix; Extracting gradient characteristics of the first net signal image matrix, and constructing a morphology and light intensity combined confidence coefficient matrix by combining pixel gray values, wherein the confidence coefficient matrix is used for representing the probability that each pixel point is an effective biological tissue fluorescence signal; carrying out weighted statistics on the first net signal image matrix and the second net signal image matrix by utilizing the confidence coefficient matrix, calculating a weighted fluorescence attenuation ratio, and reversely deducing the theoretical maximum fluorescence intensity of the sample before photobleaching according to the weighted fluorescence attenuation ratio and the sensor characteristic; And calculating the optimal exposure time finally used for formal acquisition according to the theoretical maximum fluorescence intensity, the preset target gray value and the probe exposure time in the micro-dose parameter.
2. The fluorescence immunohistochemical automatic exposure and value-taking method according to claim 1, wherein when the form and light intensity combined confidence coefficient matrix is constructed, the expression of the signal confidence coefficient weight of each pixel point is calculated as follows: ; In the formula, Representing coordinates Signal confidence weight of pixel point; Representing coordinates in a first net signal image matrix Pixel gray values at; Representing a logarithmic gain coefficient; a local gradient mode length representing the pixel point; representing the gradient penalty coefficients.
3. The method for automatically exposing and evaluating the fluorescence immunohistochemistry according to claim 2, wherein the calculating the weighted fluorescence attenuation ratio specifically comprises the steps of respectively carrying out weighted summation on a first net signal image matrix and a second net signal image matrix by using the signal confidence weight as a weighting coefficient, and dividing the weighted summation of the second net signal image matrix by the weighted summation of the first net signal image matrix to obtain the weighted fluorescence attenuation ratio.
4. The fluorescence immunohistochemical automatic exposure retrieval method according to claim 3, wherein the expression of the theoretical maximum fluorescence intensity of the back-pushed sample before the occurrence of photobleaching is as follows: ; In the formula, The theoretical maximum fluorescence intensity of the sample zero moment obtained by back-pushing is represented; Representing the currently observed valid highlight signal value; representing the weighted fluorescence decay ratio; representing a photoelectric conversion nonlinear correction factor of the sensor; Representing the photobleaching kinetics index.
5. The fluorescence immunohistochemical automatic exposure retrieval method according to claim 4, wherein the expression of calculating the optimal exposure time finally used for formal collection is as follows: ; In the formula, Indicating an optimal exposure time; representing probe exposure time when the probe frame is acquired; Representing the preset target peak gray level of the image; the theoretical maximum fluorescence intensity of the sample zero moment obtained by back-pushing is represented; representing a safety margin coefficient.
6. The method of claim 3, wherein when calculating the weighted fluorescence decay rate, if the weighted sum of the first net signal image matrix is less than a predetermined minimum energy threshold, the weighted fluorescence decay rate is directly set to 1, and the current field of view is marked as the background field of view.
7. The fluorescence immunohistochemical automatic exposure value-taking method according to claim 2, wherein the local gradient modular length is calculated by using a Sobel operator, and is obtained by calculating the sum of absolute values of gray level difference values of pixel points in the horizontal direction and the vertical direction.
8. The fluorescence immunohistochemical automatic exposure and value-taking method according to claim 1, wherein the first probe frame and the second probe frame are processed by utilizing pre-stored dark field images, specifically comprising the steps of subtracting gray values of coordinates corresponding to the dark field images from gray values of the probe frames, and comparing the result with 0 to take the maximum value so as to eliminate thermal noise of the sensor.
9. The method for automatically exposing and evaluating the fluorescent immunohistochemical image according to claim 4 is characterized in that the effective highlighting signal value is obtained by sequencing the values in the confidence coefficient matrix, selecting coordinate points corresponding to the pre-set proportion with the largest value, and calculating the average gray value of the coordinate points in the first net signal image matrix.
10. A fluorescence immunohistochemical automatic exposure value-taking system, characterized by comprising: A processor; A memory storing computer instructions for fluorescence immunohistochemical auto-exposure valuation, which when executed by the processor, cause the system to perform the fluorescence immunohistochemical auto-exposure valuation method of any one of claims 1-9.

Description

Fluorescence immunohistochemical automatic exposure value-taking method and system Technical Field The invention relates to the technical field of image data processing. More particularly, the invention relates to a fluorescence immunohistochemical automatic exposure value-taking method and a system. Background Multiple fluorescent immunohistochemical techniques can realize in-situ detection of multiple biomarkers in tumor microenvironment by marking multiple fluorescent dyes on the same tissue section, and in the digital pathological scanning process, automatic exposure is a core step for determining the accuracy of quantitative analysis of images, which directly influences the subsequent judgment of the expression level of key proteins and the reliability of diagnostic results. In the prior art, a trial-and-error feedback method is generally adopted to determine the exposure time, namely, an analysis histogram is analyzed after the initial time exposure is set, if the initial time exposure is unsuitable, the time is adjusted to be exposed again, and the process usually needs multiple times of full-power excitation to lock the optimal value. However, fluorescent dyes are very sensitive to light, each heuristic exposure results in irreversible photochemical quenching of the fluorescent molecule, and when the final defined parameters are formally imaged, the actual signal of the sample tends to be greatly attenuated, resulting in a lower quantitative score. In order to reduce photobleaching, some prior art attempts use extremely short pre-sweeps to reduce the loss of sample by greatly shortening the illumination time of the excitation light. Such improved techniques typically employ conventional methods based on full-image average gray scale or simple threshold segmentation to process the image data after the image is acquired at very short exposures, desirably by analyzing these low-dose pre-scan images to determine the brightness distribution of the sample, and then calculate the exposure parameters required for subsequent formal acquisition. However, at very short exposures, photon shot noise and sensor dark current noise in the image dominate, resulting in very low image signal-to-noise ratios. In the traditional method based on full-image average gray level or simple threshold segmentation, weak effective fluorescent signals are difficult to strip from strong background noise, so that the calculated exposure time is seriously deviated from the real requirement, and the requirement of high-precision pathological analysis cannot be met. Disclosure of Invention The invention aims to provide a fluorescence immunohistochemical automatic exposure value-taking method and a system, which can effectively avoid signal loss caused by trial-and-error exposure and accurately calculate exposure parameters in a low signal-to-noise ratio environment. To this end, the present invention provides a solution in two aspects as follows. In a first aspect, the present invention provides a fluorescence immunohistochemical automatic exposure and value-taking method, including: The method comprises the steps of controlling an excitation light source to continuously acquire a current view field with preset micro-dose parameters, obtaining a first probe frame and a second probe frame, processing the first probe frame and the second probe frame by utilizing pre-stored dark field images to obtain a first clean signal image matrix and a second clean signal image matrix after denoising, extracting gradient characteristics of the first clean signal image matrix, combining pixel gray values to construct a form and light intensity combined confidence coefficient matrix, wherein the confidence coefficient matrix is used for representing the probability that each pixel point is an effective biological tissue fluorescent signal, weighting and counting the first clean signal image matrix and the second clean signal image matrix by utilizing the confidence coefficient matrix, calculating weighted fluorescence attenuation ratio, and back-pushing the theoretical maximum fluorescence intensity of a sample before photo-bleaching according to the weighted fluorescence attenuation ratio and the sensor characteristics, and calculating the final optimal exposure time for acquisition according to the theoretical maximum fluorescence intensity, a preset target gray value and probe exposure time in the micro-dose parameters. Therefore, through double-frame micro-dose collection and subsequent mathematical reverse thrust, the bombardment of photons to a sample can be reduced to the greatest extent, the irreversible loss of signals caused by the traditional multi-test error exposure is avoided, and the protection of precious pathological samples is realized. Preferably, when the form and light intensity combined confidence coefficient matrix is constructed, the expression of the signal confidence coefficient weight of each pixel point is calculated as fol