Optimized Reference Frame Selection for Video Coding by Cloud

Bin Li,Jizheng Xu,Houqiang Li,Feng Wu

DOI: https://doi.org/10.1109/mmsp.2011.6093770

2011-01-01

Abstract:We investigate how to improve video coding efficiency via optimized reference frame selection using large-scale computation resources, e.g., a cloud. We first formulate the optimization problem for reference frame selection in video coding, which can be simplified to a manageable level. Given the maximum number of reference frames for encoding one frame, we give the upper bound of the coding efficiency on the High Efficiency Video Coding (HEVC) platform, which, although ideal, may require a huge amount of reference frame buffering at the decoder. Then we give a solution and the corresponding performance when the reference frame buffer size at the decoder is constrained. Experimental results show that when the number of reference frames is four, the proposed encoding scheme can achieve up to 16.9% bit-saving compared to HEVC, the state-of-the-art video coding system. The proposed encoding scheme is standard-compliant and can also be applied to H.264/AVC to improve coding efficiency.