X-Ray Image Contrast Enhancement Algorithms

In medicine, X-ray images are widely used for early detection and reliable diagnosis of many diseases in patients. In this case, the quality of the X-ray image is required to be high for the diagnosis to be effective. For example, the lack of sufficient contrast in the X-ray image makes it difficult for experts in the field to distinguish between the structures of the patient's internal organs. This problem can be overcome by applying contrast enhancement algorithms to the image. Today, many algorithms designed to enhance image contrast have been developed, but not all of them are equally effective for existing types of X-ray images. It depends on the specific requirements of the choice of algorithm. In this case, it is very important to evaluate the effectiveness of different contrast enhancement algorithms in a set of X-ray images and compare their results with the original images. This article analyses contrast enhancement algorithms for improving the quality of X-ray images and compares them using image quality evaluation criteria. The research work aims to determine the optimal pair of image quality evaluation criteria and contrast enhancement algorithms to ensure accurate and fast diagnosis. In the computational experiment, histogram equalization, Contrasted-limited adaptive histogram equalization (CLAHE), contrast stretching and morphological contrast enhancement algorithms were applied to 3615 human knee X-ray images and evaluated based on Peak signal noise to ratio (PSNR), Mean square error (MSE) and structure similarity index measure (SSIM) evaluation criteria. As a result, the pair formed based on the MSE criterion and the CLAHE algorithm from the contrast enhancement algorithms was determined as the most optimal pair.