Machine Learning-Based Rate Control Scheme for High Efficiency Video Coding

The up-and-coming high-efficiency video coding (HEVC) offers superior compression performance compared to the previous-generation video compression standard H.264/AVC. The rate control is the most important coding tool for compressed data transmission because of the time-varying transmission bandwidth. Based on a wide range of offline data analyses and statistics, this paper presents a rapid rate control scheme that has low computational complexity for HEVC applications within electronic devices. Experimental results demonstrate that the presented scheme can perform HEVC under time-varying bit-rate conditions.