Error resilience and concealment for Multiview video coding

This paper presents an error resilience and concealment technique in Multiview video coding. The work in this paper is aimed at making the Multiview video bitstream more robust to packet loss and improves the quality of the reconstructed video. Although error resilience in Multiview video coding has been an active research field, part of the motivation of this work is to employ data partitioning and simple frame copy error concealment in H.264/AVC and implement the two techniques into the Joint Multiview Video Coding (JMVC 8.5) reference software. This work is an attempt to illustrate the performance of the techniques implemented in the JMVC reference software in error prone channels while also maintaining an acceptable perceptual quality. The results obtained have demonstrated a considerable gain in the video quality over non-resilient coder. Thus, the technique employed gives a good performance over a packet loss network.