Image Restoration Using Graph Cuts with Adaptive Smoothing

Image Restoration is very difficult in computer vision because of the effect of complicated noise. In the paper, we propose a new image restoration algorithm using graph cuts with adaptive smoothing. First the problems can be elegantly expressed energy minimization as graph cuts and we modify the energy equation with adaptive smoothing which can solve image edge blur and discontinuity. In the iterated graph cuts the coefficient of filter can adaptively change according to gradient on each pixel. Experimental results demonstrate the approach is better for image restoration than other method.