Enhancement of Digital Images using Stochastic Resonance

Abstract

In certain non-linear dynamical systems, the addition of noise can assist the detection of a signal. In general, noise is always seen as deteriorating the system performance very effectively and is therefore normally considered as a nuisance. However, in certain non-linear systems including non-linear signal and image processing, the presence of noise can in fact enhance the weak signal strength or the low contrast images. This phenomenon is termed stochastic resonance. In this paper, we present a novel approach for contrast enhancement of low contrast images, which outperforms the other image enhancement techniques. Block diagram of our algorithm for stochastic resonance and derivation of threshold equations to be used in image contrast enhancement is also presented in this paper. The result of this processing yields a contrast- enhanced image with better visualizibility, no blocking artifacts and no other visual distortion like spot in an image. Results of our methods are compared with the enhancement methods based on histogram equalization and wavelet transform.