An information-theoretic approach for removing feedback channels in distributed video coding

Abstract

This paper presents an information-theoretic approach for estimating the compression bounds of Wyner-Ziv frames in a distributed video coding system, taking into account the amount of motion in the captured video scene, as well as the transmission channel conditions. Simulation results show a comparable behavior between the practical system performance and the theoretical limits. Therefore, based on the analytical estimations, a suitable compression rate can be predicted for each frame without the need for a feedback channel, with a minor loss in the system's rate-distortion performance compared to the classical feedback-based coding system.