New results on multiple descriptions in the Wyner-Ziv setting
IEEE Transactions on Information Theory
A cognitive approach for effective coding and transmission of 3D video
ACM Transactions on Multimedia Computing, Communications, and Applications (TOMCCAP) - Special section on ACM multimedia 2010 best paper candidates, and issue on social media
| Hi-index | 0.07 |
Most multiple description video coding (MDVC) schemes use predictive coding for inter-frame compression, suffering from the notorious drift problem, especially at high packet loss rates. In this paper, a novel MDVC scheme without any prediction loops is proposed. The new scheme is based on a code construction for the multiple description problem in the Wyner-Ziv setting [1], which is also presented in the paper. An iterative algorithm is proposed for codebook optimization. A high-rate analysis is given, showing a 3.07dB gap between the performance of the code construction and the theoretical bound. Inter-frame compression is achieved by using distributed video coding approaches instead of prediction loops. Hash check bits are transmitted instead of motion vectors. High robustness is achieved because the drift problem is completely avoided. Experiment results show that our MDVC scheme significantly outperforms the classical scheme proposed in [4] at high packet loss rates.