Enhancement X-ray image using modified dark channel prior based on Otsu segmentation

Images obtained through digital flatbed detectors in medical imaging often show insufficient quality, which poses challenges in accurate diagnosis of patients relying on these images. The proposed approach, called Modified Dark Channel Prior Based on Otsu Segmentation (MDCPOS), seeks to enhance X-ray images. This technology combines the strengths of Modified Dark Channel Prior (DCP), which identifies dark areas and enhances detail, with Otsu Segmentation, which splits the image for enhancement. By intelligently integrating these technologies, the algorithm improves the quality and clarity of X-ray images. The DCP algorithm addresses image attenuation in medical scenarios by reducing the brightness of the DCP and identifying transmission components. Otsu’s method, used in segmentation, improves performance by effectively separating bins during histogram equalization and filling black areas in X-ray images. The algorithm establishes a theoretical framework for improving X-ray images. Three quantitative metrics such as Structural Similarity Index (SSIM), Blind or Non-Referenced Image Spatial Quality Evaluator (BRISQUE), and Perceptual Image Quality Evaluator (PIQE) are used to evaluate enhanced X-ray images. The results show that the proposed method achieves the highest averages for SSIM (0.679), BRISQUE (22.664), and PIQE (34.684), quickly achieving satisfactory results that outperform other comparison methods in various critical aspects.