Scalable combined H.264/distributed video coding

Abstract

PurposeThe first purpose of this paper is to propose a scalable video coding scheme providing flexibility in video transmission, especially under wireless environment. The second purpose is to analyze the problem of lengthening the key frame interval in distributed video coding (DVC), and propose an approach to improve the rate‐distortion (RD) performance of DVC for long group‐of‐frames (GOF) size.Design/methodology/approachIn the proposed scheme, a base layer is first obtained from an H.264 coder. When a DVC coder is then used to code the enhancement layer, information in processing the base layer is extracted and analyzed to make multiple side‐information available and reduce error accumulation for DVC coding, thus further improving the performance of the DVC coder.FindingsBy dividing video into base and enhancement layers, the combined video coding architecture enables a flexible video transmission. In addition, several methods are used to improve the RD performance in DVC coding. Simulation shows that the proposed scheme outperforms non‐scalable DVC for long GOF size.Originality/valuePrediction from the decoding loop in base layer encoder largely reduces enhancement layer spatial redundancy. Multiple side‐information provides better estimation for DVC reconstruction. Long prediction loop is more reliable because error accumulation is effectively compensated.