CN-121998886-A - OCT image self-adaptive contrast enhancement method

CN121998886ACN 121998886 ACN121998886 ACN 121998886ACN-121998886-A

Abstract

The application relates to the technical field of image processing, and provides an OCT image self-adaptive contrast enhancement method, which comprises the steps of acquiring an OCT image, normalizing the OCT image to a preset gray scale interval, and obtaining a normalized image; and determining a target gray upper limit value for eliminating an overexposure region in the OCT image and a target gray lower limit value for eliminating a background noise region in the OCT image according to gray distribution statistical information of the normalized image, constructing an enhancement function based on a Sigmoid function, wherein a definition domain is limited by the target gray upper limit value and the target gray lower limit value, outputting values for representing enhanced pixel gray, and applying the enhancement function to the normalized image pixel by pixel to generate an enhanced image. According to the method, a smooth Sigmoid curve is adopted while the intermediate gray contrast is stretched, the definition domain is limited to an effective signal interval, dark part noise and bright part details are protected, and therefore the enhancement effect and the practicability are both considered.

Inventors

JIANG CHONG
LIU CHUNYAN
TANG XU
ZHAO ZHENDONG
ZHOU TONG

Assignees

江苏富翰医疗产业发展有限公司

Dates

Publication Date: 20260508
Application Date: 20260409

Claims (10)

1. An OCT image adaptive contrast enhancement method, comprising: Acquiring an OCT image, normalizing the OCT image to a preset gray scale interval, and obtaining a normalized image; Determining a target gray upper limit value and a target gray lower limit value according to the gray distribution statistical information of the normalized image, wherein the target gray upper limit value is used for eliminating an overexposure region in the OCT image, and the target gray lower limit value is used for eliminating a background noise region in the OCT image; Constructing an enhancement function, wherein the enhancement function is a nonlinear mapping function based on a Sigmoid function, the definition domain of the enhancement function is limited by the target gray upper limit value and the target gray lower limit value, and the output value of the enhancement function is used for representing the enhanced pixel gray; And applying the enhancement function to the normalized image pixel by pixel to generate an enhanced image.
2. The OCT image adaptive contrast enhancement method of claim 1, wherein determining a target gray upper value and a target gray lower value from gray distribution statistics of the normalized image comprises: calculating a gray level histogram of the normalized image; counting a cumulative distribution function according to the gray level histogram; when the cumulative distribution function reaches a first preset threshold value, determining a corresponding gray value as the target gray lower limit value; And when the cumulative distribution function reaches a second preset threshold value, determining the corresponding gray value as the target gray upper limit value.
3. The OCT image adaptive contrast enhancement method of claim 2, wherein the first preset threshold is 0.5 and the second preset threshold is any one of 0.99-0.995.
4. The OCT image adaptive contrast enhancement method of claim 1, wherein after the determining the target gray upper limit value and the target gray lower limit value, the method further comprises: acquiring a difference value between the target gray upper limit value and the target gray lower limit value; and when the difference value is smaller than a preset width threshold value, marking the quality of the OCT image as an abnormal state, and when the quality of the OCT image is in the abnormal state, replacing the target gray upper limit value and the target gray lower limit value with preset default intervals.
5. The OCT image adaptive contrast enhancement method of claim 1, wherein after the determining the target gray upper limit value and the target gray lower limit value, the method further comprises: obtaining the maximum gray value of the normalized image; when the maximum gray value is smaller than a preset full black threshold value, judging that the OCT image is a full black image; and outputting the normalized image as the enhanced image.
6. The OCT image adaptive contrast enhancement method of claim 1, wherein the constructing an enhancement function comprises: obtaining a standard Sigmoid function; The function value of the standard Sigmoid function at the zero point is adjusted to be zero, and the Sigmoid function with zero point return is obtained; And normalizing the function value of the zero-return Sigmoid function at the target gray upper limit value to unit amplitude to obtain the enhancement function.
7. The OCT image adaptive contrast enhancement method of claim 6, wherein the adjusting the function value of the standard Sigmoid function at the zero point to zero results in a zero-point zeroed Sigmoid function, comprising: Acquiring a midpoint value of the target gray upper limit value and the target gray lower limit value; taking the midpoint value as an independent variable offset, and subtracting the midpoint value from the independent variable of the standard Sigmoid function to obtain a Sigmoid function after the independent variable offset; And acquiring a function value of the independent variable offset Sigmoid function when the independent variable is zero as a zero point offset, and subtracting the zero point offset from the independent variable offset Sigmoid function to obtain a zero point zeroed Sigmoid function.
8. The OCT image adaptive contrast enhancement method of claim 6, wherein normalizing the function value of the zero-zeroed Sigmoid function at the target gray upper limit to a unit magnitude results in the enhancement function, comprising: Calculating the function value of the zero-return Sigmoid function at the upper limit value of the target gray scale as an amplitude normalization factor; Dividing the zero-return Sigmoid function by the amplitude normalization factor to obtain an intermediate function with a value range of 0-1 closed interval; Acquiring a preset output gray minimum value and a preset output gray maximum value; And after summing the intermediate function and the output gray minimum value, multiplying the difference between the output gray maximum value and the output gray minimum value to obtain an enhancement function with a value range from the output gray minimum value to the output gray maximum value closed interval.
9. The OCT image adaptive contrast enhancement method of claim 6, wherein the standard Sigmoid function comprises a steepness parameter for controlling a steepness of the enhancement function between the target gray lower limit value and the target gray upper limit value, wherein the steepness parameter has a negative correlation with a difference between the target gray upper limit value and the target gray lower limit value.
10. The OCT image adaptive contrast enhancement method of claim 9, wherein the sharpness parameter is determined by: acquiring a difference value between the target gray upper limit value and the target gray lower limit value; Acquiring a preset constant; and determining the ratio of the preset constant to the difference value as the steepness parameter.

Description

OCT image self-adaptive contrast enhancement method Technical Field The application relates to the technical field of image processing, in particular to an OCT image self-adaptive contrast enhancement method. Background An optical coherence tomography (Optical Coherence Tomography, OCT) technology acquires a tomographic image of a tissue such as retina through a low coherence light interference principle, and an imaging system formed by the OCT technology consists of a light source, an interferometer and a detector and outputs a gray level image reflecting the backward scattering intensity of the tissue. In clinical diagnosis, doctors expect images to clearly present weak structural differences of tissues of each layer while suppressing background noise, which makes clear demands on the self-adaption capability and fidelity of a contrast enhancement algorithm. The partial image enhancement technology comprises a global mapping method based on power law transformation, applying unified nonlinear transformation to the whole image by adopting preset fixed parameters, constructing a deep neural network based on a data-driven enhancement scheme, and training an end-to-end mapping model from an original image to an enhancement result by utilizing a large-scale pairing data set. The global mapping method is fixed in parameters and irrelevant to image content, so that effective signals and background noise are difficult to distinguish, dark part noise amplification or bright part detail loss is easy to occur when middle gray scale is stretched, the data driving method relies on high-quality pairing training data, the mapping rule is hidden in network parameters and does not have explicit mathematical expression, and the calculation amount of the reasoning process is large. In summary, it is difficult to achieve both enhancement effect and practicality in the image enhancement technique. Disclosure of Invention The application provides an OCT image self-adaptive contrast enhancement method, which aims to solve the problem that the enhancement effect and the practicability of an image enhancement technology are difficult to be compatible. The application provides an OCT image self-adaptive contrast enhancement method, which comprises the following steps: Acquiring an OCT image, normalizing the OCT image to a preset gray scale interval, and obtaining a normalized image; Determining a target gray upper limit value and a target gray lower limit value according to the gray distribution statistical information of the normalized image, wherein the target gray upper limit value is used for eliminating an overexposure region in the OCT image, and the target gray lower limit value is used for eliminating a background noise region in the OCT image; Constructing an enhancement function, wherein the enhancement function is a nonlinear mapping function based on a Sigmoid function, the definition domain of the enhancement function is limited by the target gray upper limit value and the target gray lower limit value, and the output value of the enhancement function is used for representing the enhanced pixel gray; And applying the enhancement function to the normalized image pixel by pixel to generate an enhanced image. In some possible embodiments, the determining the target gray upper limit value and the target gray lower limit value according to the gray distribution statistical information of the normalized image includes: calculating a gray level histogram of the normalized image; counting a cumulative distribution function according to the gray level histogram; when the cumulative distribution function reaches a first preset threshold value, determining a corresponding gray value as the target gray lower limit value; And when the cumulative distribution function reaches a second preset threshold value, determining the corresponding gray value as the target gray upper limit value. And determining the upper limit and the lower limit by using a cumulative distribution function of the gray level histogram, so that the enhancement zone is completely based on the content of the image, and realizing self-adaptive distinguishing of the effective signal and the background noise. In some possible embodiments, the first preset threshold is 0.5, and the second preset threshold is any one of 0.99-0.995, so as to separate the background noise and the overexposure region, and ensure that the enhancement function only acts on the effective signal interval. In some possible embodiments, after the determining the target gray upper limit value and the target gray lower limit value, the method further includes: acquiring a difference value between the target gray upper limit value and the target gray lower limit value; And when the difference value is smaller than a preset width threshold value, marking the quality of the OCT image as an abnormal state, and when the abnormal state is the abnormal state, replacing the target gray upper limit valu