Image representation and enhancement on elastic manifolds
Vadim Ratner, Yehoshua Y. Josh Zeevi · 2008
Enhancing detail and contours of objects in low resolution noisy natural images is still one of the most outstanding challenges in image processing. This study is devoted to a new approach to adaptive image enhancement, based on the application of a damped elastic deformation process to single images. The proposed approach results in a Telegraph-Diffusion (TeD) denoising-and-sharpening filtering scheme which enhances fine details in images. Three efficient numerical schemes are presented. Advantages of the algorithm are discussed with reference to computational results.