KSVD-Based Multiple Description Image Coding

Abstract

In this paper, we present a new multiple description coding scheme, which is based on a sparse dictionary training method called K singular value decomposition (KSVD). In the proposed scheme, each description encodes one source subset with a small quantization stepsize, and other subsets are predictively coded with a large quantization stepsize. The source processed by the KSVD becomes sparse, which can improve the coding efficiency. The proposed scheme is then applied to lapped transform-based multiple description image coding. Finally, image coding results show that the proposed scheme achieves a better performance than the current state-of-the-art multiple description coding methods.

Keywords

• K singular value decomposition (KSVD)
• multiple description coding
• sparse representation