HEVC Rate Control Algorithm Based on Predictive Weight

Abstract

Aiming at the shortcomings of HEVC (High Efficiency Video Coding) rate control algorithm using only the information of inter-coded units without using the information of the intracoded units in the frame to allocate the target bit and update the model parameters, this article proposed a LCU (Largest Coding Unit) layer rate control algorithm based on predictive weight. The proposed algorithm makes full use of the information of both inter-coded units and intra-coded units in the frame, and utilizes a target bit allocation method based on the prediction weight and a model parameter updating method based on the prediction weight which can accurately achieve the target output bit rate, and improve video quality while ensuring the smooth stability of video quality. The experimental results show that in the premise of the same rate control error, BD-PSNR (Bjøntegaard delta peak signal-to-noise rate) of improved algorithm is improved by 0.059dB on average on Low- Delay configuration, and 0.040dB on average higher on Random-Access configuration, and the video quality PSNR(Peak Signal to Noise Ratio)of improved algorithm is less volatile.