Backward channel aware Wyner–Ziv video coding: A study of complexity, rate, and distortion tradeoff

Abstract

Wyner–Ziv video coding is a coding method that exploits source statistics at the decoder. Many Wyner–Ziv video coding schemes encode a video sequence into two types of frames, key frames and Wyner–Ziv frames, and rely on the frequent use of INTRA frames as key frames such that the decoder can derive sufficiently accurate side information. While the use of INTRA frame as key frames keeps the encoding complexity low, it also leads to much lower coding efficiency compared to state-of-the-art predictive video coding. In this paper we present a Wyner–Ziv video coding scheme that uses backward channel aware motion estimation to encode the key frames, where motion estimation is performed at the decoder and motion information is sent back to the encoder. Two motion vector selection methods are proposed and we present a model to study the complexity–rate–distortion tradeoff in such coding schemes.