Hierarchical Complexity Control of HEVC for Live Video Encoding

As the latest video coding standard, High Efficiency Video Coding (HEVC) tremendously improves the encoding efficiency compared with the preceding H.264/AVC standard, but at the cost of higher encoding complexity. This huge encoding complexity makes the implementation of HEVC intractable on live videos. For coping with this problem, two major challenges need to be solved: 1) How to accurately reduce the encoding complexity to achieve the target complexity? and 2) How to maintain the video quality after encoding complexity reduction? To solve these two challenges, we propose, in this paper, a hierarchical complexity control approach of HEVC. For the first goal, the complexity control is implemented in two levels to assure the control accuracy. In the largest coding unit (LCU) level, we adjust the maximum depths of LCUs in a frame to reduce the encoding complexity to the target. Since each frame has numerous LCUs, and each LCU can choose its maximum depth from one of the four maximum depths, the large degree of freedom contributes to the high control accuracy. However, there may be still some errors. These errors can be compensated in the frame level by a proposed frame level complexity control algorithm. For the second goal, one objective weight map and one subjective weight map are proposed to use in the process of complexity control to keep the objective and subjective video quality simultaneously. Finally, The experimental results show that our approach outperforms other state-of-the-art approaches, in terms of both control accuracy and video quality.