Distributed monoview and multiview video coding

Video compression has now been recast into a distributed source coding framework, paving the way to distributed video coding (DVC) systems targeted at low coding complexity and error resilience for today's mobile device needs, where video will be encoded using low cost devices and decoded on powerful platforms. For monoview systems, there are two DVC architectures. The first applies the Wyner-Ziv (WV) coding principles to either the pixel domain or in the transform domain. The spatial redundancy within images is exploited through transposing WZ coding from the pixel. Another architecture is about taking the WZ coding decision at the frame level. The sequence is thus structured into groups of pictures in which selected frames are intercoded and intermediate frames are WZ coded. Each WZ frame is encoded independently of the other frames. DVC has many potential applications, such as more flexible coder/decoder complexity balancing, increased error resilience and the capability to exploit inter-view correlation, with limited inter-camera communication in multiview set-ups. It provides benefits in layered representations, with increased error resilience and independence between codecs used in different layers. It is just a matter of maturity of bringing monoview and multiview DVC closer to predictive coding.