Enhanced Fractal-Wavelet Image Denoising

Abstract

This paper presents an enhanced fractal-wavelet image denoising (EFWID) algorithm by adopting quadratic function for fractal scale prediction in wavelet domain. It consists in an extension of an enhanced fractal image denoising (EFID) algorithm proposed by the authors for denoising the images degraded by additive white Gaussian noise (AWGN). In terms of the quality of the fractal-wavelet representation of the noiseless images, the enhanced fractal-wavelet coding (EFWC) generally performs better than the traditional fractal-wavelet image coding algorithm. Based on the enhanced fractal-wavelet coding, the improved denoising method is implemented by estimating the fractal scale coefficients of the quadratic function of the noiseless image in wavelet domain from its noisy observation, and hence to remove the noise in the decoding procedure. Experimental results show that, compared with some other traditional image denoising algorithms, the EFWID method can improve the quality of the restored image efficiently.