A Novel Image Restoration Algorithm Using NSCT and Pixel Level Fusion.

This paper aims at providing an image enhancement algorithm for digital images especially corrupted by various noises. Image restoration is a technique used to construct a composite image containing mutual and significant information from noisy source image. The novelty of this paper is implemented via applying contourlet transform as well as directional filter bank. The source image is decomposed into base and detail layers, firstly. Then, the detail layer is further decomposed by DFT filter banks to pass away noises. The denoisy detail layer is appended to the base layer of the corresponding source image to obtain noise-free source images. Finally, all the noise-free source images are fused using average fusion rule. Resulted images using our algorithm are compared with state of the art denoisy methods to show the effectiveness of the proposed method. Our proposed method outperforms the state of the art fusion method. [ABSTRACT FROM AUTHOR]