Fusion-based multiview distributed video coding

摘要

In this paper, we introduce a scheme for coding video surveillance camera networks. It is based on multiview and Distributed Video Coding (DVC). DVC is known for its low complexity encoders. This makes it very interesting for a wide range of applications, in particular video surveillance. More specifically, we introduce a new fusion technique between temporal side information and homography-based side information that improves the rate-distortion performance of the DVC compression. Finally, we introduce a new scheme with a pure Wyner-Ziv camera, which encodes all its frames as Wyner-Ziv frames.