Human visual system aware decoding strategies for Distributed Video Coding
Sören Sofke, Ralph Hänsel, Erika Müller · 2009
Distributed Video Coding (DVC) is a recent paradigm that offers emerging capabilities in contradiction to the established conventional video coding techniques. Based on the Slepian-Wolf (SW) and Wyner-Ziv (WZ) theorems, a DVC system has an outstanding low encoding complexity, by shifting the computational complex process of correlation exploration to the decoder. The fundamental DVC architecture is organized to reach a competitive Rate-Distortion (RD) performance in terms of PSNR, despite its low correlation with the human visual system (HSV). In contrast, this paper addresses the problem by proposing three image processing tools for exploiting spatio-temporal correlations to reduce the perceptual distortion of WZ frames. The proposed WZ pixel domain framework offers a comparable RD-performance referring to H.264 AVC intra coding.