Chest X-ray images often suffer from low contrast, noise, and loss of local details, which can hinder the accurate identification of bones and soft tissues and affect subsequent automated diagnosis and clinical interpretation. To address these challenges, recent methods have explored both traditional contrast enhancement and deep learning-based strategies. However, most existing methods focus on a single degradation factor, such as contrast enhancement, without explicitly considering region-specific visibility or structure-aware enhancement. In this work, we propose an adaptive enhancement approach for chest X-ray images. Our framework integrates tissue attenuation visibility enhancement, linear transformation-based contrast adjustment, and a perceptual fusion module to emphasize anatomical features, local details, and global brightness. This design aims to suppress soft-tissue interference, improve visibility of bones and improve the visualization of essential anatomical components, while maintaining consistent image quality across regions with varying exposure. Importantly, our approach does not require extensive training and is interpretable. Experiments on benchmark datasets demonstrate that the method improves image contrast, sharpness, and visibility of anatomical structures, showing promising results compared with 12 existing traditional and deep learning-based enhancement methods.
Zhao et al. (Thu,) studied this question.