CN-121788392-B - X-ray image enhancement method, system and storage medium

CN121788392BCN 121788392 BCN121788392 BCN 121788392BCN-121788392-B

Abstract

The invention belongs to the technical field of medical image processing, and particularly relates to an X-ray image enhancement method, an X-ray image enhancement system and a storage medium. The method comprises the steps of inputting an original X-ray image, re-applying window width and window level, performing noise reduction treatment, copying the X-ray image into a normal tissue image layer, a cancer tissue image layer and a calcification point image layer, adjusting brightness and contrast of each image layer by using a gamma adjustment treatment method, performing image enhancement and Retinex enhancement, obtaining three image layers subjected to image enhancement treatment and a normal tissue image layer subjected to Retinex enhancement treatment, and performing Laplace pyramid fusion to obtain an enhanced image. The invention also provides an X-ray image enhancement system for realizing the method. The invention can obtain excellent image enhancement effect on the premise of excessively amplifying or enhancing contrast, provides support for rapid and accurate diagnosis of doctors, and has good application prospect.

Inventors

BU HONG
LI JING
CHEN JIE
ZHANG GUANGYA
ZHANG ZHANG
DENG JIE
YI YUHAO
YANG YU

Assignees

四川大学华西医院
中国工程物理研究院流体物理研究所

Dates

Publication Date: 20260505
Application Date: 20260305

Claims (6)

1. An X-ray image enhancement method, comprising the steps of: inputting an original X-ray image, and re-applying window width and window level to the original X-ray image; Performing noise reduction processing on the X-ray image subjected to window width and window level reapplication processing; Copying the X-ray image subjected to noise reduction treatment into a normal tissue layer, a cancer tissue layer and a calcification point diagram layer, and adjusting the brightness and the contrast of each layer by using a gamma adjustment treatment method, wherein the gamma value of the normal tissue layer is 3, the gamma value of the cancer tissue layer is 15, and the gamma value of the calcification point diagram layer is 20; Respectively carrying out image enhancement on the three image layers subjected to gamma adjustment treatment; Carrying out Retinex enhancement on the normal tissue layer subjected to gamma adjustment treatment; acquiring three image layers subjected to image enhancement processing and a normal tissue image layer subjected to Retinex enhancement processing, and carrying out Laplacian pyramid fusion on the four image layers to obtain an enhanced image; the image enhancement processing method comprises the following steps: Local contrast enhancement, namely, respectively carrying out image enhancement on the three image layers by adopting a bidirectional sliding window contrast limiting self-adaptive histogram equalization method; image detail enhancement, namely processing three image layers by adopting a Sobel self-adaptive sharpening mask; in the local contrast enhancement, three layers are respectively enhanced by the following steps: s310, image is processed Divided into sizes of Is used to determine the non-overlapping local blocks of (1), for each block Calculating gray level histogram The calculation formula is as follows: (3) Wherein, the The gray level is represented by a gray scale, As a dirac function, x and y are pixel point coordinates respectively; s320, taking the output of S310 as input, using contrast limitation to obtain a histogram Is well above a threshold value Is uniformly distributed to obtain a limited histogram ; (4) Wherein L is 65536; S330, taking the output of S320 as input, calculating Is a cumulative distribution function of (2) And normalize it to a range Obtaining normalized cumulative distribution function , (5) Wherein, the Is that Is set to be a minimum value of (c), Is that Is the maximum value of (2); S340, taking the output of S330 as input, using step size And Enhancing the image for each pixel within the window Find four neighboring blocks where it is located And calculates interpolation weights And , (6) Wherein, the And Is the initial coordinate of the sliding window in the horizontal and vertical directions, mod is the modulo operation; S350, obtaining enhanced pixel values through bilinear interpolation Finally combining all pixel values to generate an enhanced image The calculation process of the bilinear interpolation is shown in the formula (7): (7); The steps of Sobel self-adaptive unsharp mask processing comprise: Computing average gradients of input images by Sobel operator Adjust standard deviation And enhanced strength As shown in the formula (8), (8) Wherein the method comprises the steps of Is that Is set at the maximum value of (c), Is that Is the minimum of (2); The Retinex enhancement step includes: the multi-scale decomposition is carried out by adopting Gaussian blur with standard deviation of g=5, 10 and 20 respectively, the specific decomposition formula is as follows, (9) Wherein R (g) is a decomposition result, log is a logarithmic operation, image is a normal tissue layer, blurred is a Gaussian blur image adopting g standard deviation for image; Adding equal weights to the images after decomposing the three standard deviation Gaussian blur images, and integrating multi-scale details; the step of laplacian pyramid fusion comprises the following steps: The three image layers after the image enhancement processing are respectively marked as normal tissue image layers Cancer tissue layer Calcification dot pattern layer The normal tissue layer subjected to Retinex enhancement treatment is recorded as ; Will be And Fusing, wherein the fusion ratio is 0.6:0.4, and obtaining ; Will be And Fusing at a fusion ratio of 0.8:0.2 to obtain ; Will be And Fusing to obtain an enhanced image, wherein the fusion ratio is 0.8:0.2 。
2. The method for enhancing an X-ray image according to claim 1, wherein said re-application of window width and level comprises the steps of: S110, performing inverse operation on the two-dimensional X-ray photographic image, and counting pixel value distribution, stretching the input two-dimensional image into a one-dimensional sequence, calculating histogram statistics and cumulative distribution on the sequence, and normalizing the one-dimensional sequence by using the minimum and maximum pixel values; S120, obtaining the output of S110 as input, finding the interval range of 95% pixel [ , Calculating window width and window level, the specific calculation is shown in formula (1), (1) Wherein, the And Respectively representing the window width and the window level, And Respectively representing an upper limit value and a lower limit value of a 95% pixel interval; S130, obtaining the output of the S120 as input, and calculating the upper limit value and the lower limit value of a window as shown in a formula (2); (2) Wherein, the The upper limit value of the window is indicated, Representing the lower limit of the window.
3. The method for enhancing an X-ray image according to claim 1, wherein said noise reduction processing is Gaussian blur.
4. The method of claim 1, wherein the gamma adjustment is performed by normalizing the pixel values of the image to [0,1] using maximum pixel value normalization and adjusting the brightness of the image using gamma adjustment.
5. An X-ray image enhancement system for implementing the X-ray image enhancement method of any one of claims 1-4, comprising: an input module configured to input an original X-ray image; a window width and level application module configured to re-apply a window width and level to the original X-ray image; the noise reduction module is configured to perform noise reduction processing on the X-ray image processed by the window width and level application module; The gamma adjusting module is configured to copy the X-ray image processed by the noise reducing module into a normal tissue layer, a cancer tissue layer and a calcification point layer, and adjust the brightness and the contrast of each layer by using a gamma adjusting processing method; The enhancement module is configured to respectively carry out image enhancement on the three image layers processed by the gamma adjustment module; The Retinex shading module is configured to carry out Retinex enhancement on the normal organization layers processed by the gamma adjustment module; And the fusion module is configured to acquire three layers processed by the enhancement module and a normal organization layer processed by the Retinex shading module, and fuse the four layers by using a Laplacian pyramid to obtain an enhanced image.
6. A computer-readable storage medium, having stored thereon a computer program for implementing the X-ray image enhancement method according to any one of claims 1 to 4, or a computer program for implementing the X-ray image enhancement system according to claim 5.

Description

X-ray image enhancement method, system and storage medium Technical Field The invention belongs to the technical field of medical image processing, and particularly relates to an X-ray image enhancement method, an X-ray image enhancement system and a storage medium. Background Cancer is one of the leading causes of death worldwide, and its pathogenesis is complex involving multiple factors such as gene mutation, uncontrolled cell proliferation and immune escape. The latency period of the cancer is long, early asymptomatic is difficult to find, the optimal treatment time is missed, and the cancer is the root cause of high and persistent cancer mortality. Early treatment was found to be the most effective means of coping with cancer. X-ray imaging technology has become an indispensable tool for medical diagnosis since discovery, especially early cancer diagnosis. The method generates a gray image reflecting the internal structure by penetrating human tissue and capturing the absorption difference of different tissues on X-rays. However, the quality of X-ray images is often limited by a number of factors, such as noise, low contrast, edge blurring, and physical limitations of the device itself. These problems can lead to doctors missing important details during the diagnosis and even to misdiagnosis. Calcification foci, a potentially important imaging marker for cancer, was identified early on primarily by means of X-ray photography. The traditional X-ray can emit rays penetrating through tissues, so that tissues with different densities can be imaged on a film, and a tiny calcification focus can be in a bright point or granular compact image in an image due to the difference of the radiation absorption degree of surrounding soft tissues. However, there are a number of difficulties in determining the nature of calcification lesions solely by means of radiography. On the one hand, benign lesions such as post-traumatic repair and the like may be calcified, and the calcification related to malignant tumors may overlap in morphology and distribution, so that the calcification is easily confused when being interpreted by naked eyes of doctors, and misdiagnosis or overdiagnosis is caused. On the other hand, the two-dimensional imaging characteristic of the X-rays makes it difficult for some calcification foci to be accurately positioned in tissues overlapped in front and back, and influences subsequent further diagnosis and treatment planning. The X-ray image enhancement technique aims to improve the quality of the X-ray image, making it clearer and easier to diagnose. The X-ray image enhancement technique highlights the lesion area and the microstructure by means of contrast adjustment, noise suppression, and the like. In the prior art, common methods include histogram equalization and filters. The technology not only improves the visual effect of the image, but also assists doctors to identify diseases more accurately, and has important application value in medical image diagnosis. However, these existing methods are still deficient in enhancing the effect of the X-ray image. Specifically, there are three key issues that need to be resolved: (1) The contrast is often adjusted in existing X-ray image enhancement techniques, however, excessive magnification or contrast enhancement operations may introduce artifacts. These artifacts are not true calcification focus information, but easily interfere with the physician's vision, causing confusion in the judgment, and even misregarding the artifacts as lesion features, resulting in misdiagnosis. (2) The tissue structures of different individuals are complex and various, the calcified foci are quite different in morphology and distribution, the excessive amplification or contrast enhancement is difficult to accurately adapt to all conditions, the phenomenon of misjudgment or misjudgment of the calcified foci occurs, and uncertainty is brought to clinical decisions. (3) The prior art generally focuses on enhancing calcification foci, however, it is equally important to enhance other areas of the X-ray image (e.g. tissue), which makes it hierarchical and clearer, and is also an important way to help diagnose cancer. In view of the foregoing, there is still a need in the art to develop new X-ray image enhancement techniques that achieve good enhancement of the overall X-ray image without excessive amplification or contrast enhancement. Disclosure of Invention In view of the foregoing problems with the prior art, the present invention provides an X-ray image enhancement method, system, and storage medium. An X-ray image enhancement method comprising the steps of: inputting an original X-ray image, and re-applying window width and window level to the original X-ray image; Performing noise reduction processing on the X-ray image subjected to window width and window level reapplication processing; Copying the X-ray image subjected to noise reduction treatment into a normal tiss