Selective image smoothing and feature enhancement using modified shock filters

Shock filters are widely used for image enhancement and deblurring. These filters make use of nonlinear hyperbolic Partial Differential Equations (PDEs) in order to sharpen the edges. However, in many practical cases images are corrupted by noise and other kind of degradations. Convensional shock filters are not suitable in such cases as they enhance the noise present in the image. Hence, the idea of combining shock filters with diffusion yield good results. In this paper we propose a modified "diffusion coupled shock filter." The proposed method makes use of an 'adaptive diffusion term' which limits the extent of smoothing on important edges making them more sharper. The experimental results demonstrate the efficiency of the proposed method to control diffusion and to make the reconstruction more reliable.