Cross-layer error resilient mechanism in scalable video coding

Scalable video coding (SVC), which is the scalable extension of H.264/AVC, is designed to transmission over heterogeneous networks and provide scalabilities and high compression ratio. However, such kinds of compressed videos are sensitive to transmission error, resulting in quality degradation seriously. In this paper, a cross-layer error resilient method is proposed to prevent transmission error in SVC. In the proposed algorithm, the reversible data embedding technique is adopted to hide the essential information of lower layer such as IntraBL and the predicted motion vector (MV) for the reconstruction of higher layer. Experimental results show that the proposed method can provide better PSNR performance than frame copy by 6.76 dB on average.