Adaptive averaging of multiresolution bilateral and median filtering for image denoising

Abstract

In this paper the approach of denoising is solved by a new hybrid technique, which associates the different denoising methods. Wavelet threshold, bilateral filtering and median filtering are the three different filters employed in our hybrid technique. While doing image denoising, Spatial averaging and Edge preservation are the two different characteristics which will be considered and obviously each filter will have its own characteristics. The bilateral filter does the spatial averaging without smoothing the edges whereas a wavelet threshold filter removes the noise by removing the high frequency components with relatively lesser edge preservation. A median filter removes the noise with very good edge preservation. No filter gives the robust result in terms of both spatial averaging and edge preservation. Our new method proposes an adaptive combination of these three filtering methods, which outperforms the different denoising methods in terms of PSNR and edge preservation. Simulated experimental results were given for the proposed method.