Interleaving-based error concealment for scalable video coding system

Abstract

Scalable video coding (SVC) is desirable for video communication in heterogeneous environments where end-users retain different types of terminals. Because motion estimation and compensation are used in SVC to reduce redundancy between frames, the visual quality degrades due to the errors of transmission channels and the error propagation across multiple frames. In order to decrease error propagation and enhance visual quality of reconstructed frames, error concealment method is used to estimate the erroneous motion vectors by exploiting the spatial and temporal correlation of video frames in different layers. Compared with two existing error concealment methods, an improved error concealment method is proposed with incorporation of interleaving technique. A spatial-temporal scalable video coding system is implemented in two layers. Especially, the impact of burst errors on motion vectors are investigated in this paper. Experimental results verify that the proposed error concealment method with interleaving technique outperforms the existing methods.