A Frame Loss Error Concealment Scheme for Multi-view Video Coding

Robust multi-view video transmission over error-prone networks is a technical challenge problem for multi-view video coding (MVC). In this paper, a frame loss error concealment scheme is proposed for MVC to address the challenging problem at the decoder side. Different from traditional motion information derivation in temporal domain for single-view video coding, two novel information derivation schemes, motion information derivation in inter-view domain and disparity information derivation in temporal domain, are proposed in this paper. Each block of lost frames can be reconstructed from its temporal reference frame and its interview reference frame by merging different concealment results under the three prediction information derivations. Simulation results show that the proposed scheme can improve frame loss error concealment performance for MVC.