Distributed Video Coding

Distributed video coding is a new compression paradigm based on two key Information Theory results: the Slepian-Wolf and Wyner-Ziv theorems. Those results suggest that it is possible to compress two statistically dependent signals in a distributed way (separate encoding, jointly decoding) using a rate similar to that used in a system where the signals are encoded and decoded together, i.e. like in traditional video coding schemes. This theoretical framework is referred as lossy source coding with side information at the decoder (Wyner-Ziv) and enables to shift the complexity from the encoder to the decoder.

Future work should be posted here.

Improving the Quality of the Side Information:

The IST wyner-ziv video codec generates the side information by using motion compensated frame interpolation algorithms at the decoder. Besides forward and bidirectional motion estimation, a spatial motion smoothing algorithm to eliminate motion outliers is proposed (see Figure bellow). This allows significant improvements in the rate-distortion (RD) performance without sacrificing the encoder complexity.

J. Ascenso, C. Brites, F. Pereira, "Improving Frame Interpolation with Spatial Motion Smoothing for Pixel Domain Distributed Video Coding", 5th EURASIP Conference on Speech and Image Processing, Multimedia Communications and Services, Slovak Republic, July 2005. (paper) (presentation)

J. Ascenso, C. Brites, F. Pereira, "Motion Compensated Refinement for Low Complexity Pixel Based Distributed Video Coding", IEEE International Conference on Advanced Video and Signal Based Surveillance, Como, Italy, September 2005. (paper) (poster)

L. Natário, C. Brites, J. Ascenso, F. Pereira, “Side Information Extrapolation for Low-Delay Pixel-Domain Distributed Video Coding”, International Workshop on Very Low Bitrate Video, Sardinia, Italy, September 2005. (paper)

Intra mode decision for Wyner-Ziv Video Coding:

The quality of the motion-interpolation estimate is not constant across the whole frame. In this work is assumed that covered/uncovered regions, illumination changes and camera noise do affect significantly the quality of the motion interpolation image and therefore is necessary to encode some regions in intra-frame mode when correlation with the side information is weak. The Figure bellow illustrates the mode decision map performed at the encoder (low-complexity decision). Greyed areas refer to Wyner-Ziv blocks.

A. Trapanese, M. Tagliasacchi, S. Tubaro, J. Ascenso, C. Brites, F. Pereira, "Embedding a Block-based Intra Mode in Frame-based Pixel Domain Wyner-Ziv Video Coding", International Workshop on Very Low Bitrate Video, Sardinia, Italy, September 2005. (paper)