A Wavelet-Domain Bayesian Image Enhancement Approach in Distributed Video Sensor Network

The aim of image enhancement over distributed video sensor network (DVSN) is to fuse multiple images acquired by a set of video sensors to produce a single image with more visual information. For that, a wavelet-domain Bayesian approach is proposed in this paper. The proposed approach estimates the wavelet coefficients of the desired image by exploiting both the information provided by multiple images observed in DVSN as well as the prior image information that is imposed by the proposed three-component exponential mixture (TCEM) model. Extensive experiments are conducted to demonstrate the superior performance of the proposed approach.