Reference frame selection algorithm of HEVC encoder for low power video device

Abstract

Modern video applications on mobile devices use higher resolutions and frame rates in order to offer better visual quality. Therefore, the power consumption reduction of the decoder in the mobile device is more challenging. In this paper, the reference frames selection algorithm of HEVC encoder for power consumption reduction on the decoder side is proposed. In particular, the power consumption of decoder for random access mode is significant compared with other coding mode. Firstly, the evaluation result show that when the number of group of picture and reference frames are varied on some conditions in random access mode. Next, based on the evaluation results, the proposed algorithm adaptively select the optimal number of reference frames in terms of the relationship of the prediction mode, coding unit size. The proposed algorithm is evaluated by the reference software of HEVC. The simulation results show that the proposed algorithm can reduce over 40% of the reference frames comparing to the original HEVC algorithm.