Sample Length-adaptive Prediction Scheme for Parameter Estimation in Multi-dimensional Rate Control

Abstract

Multi- rate control technology has important applications in rate control and video transmission. However, most existing methods do not give measure to improve the accuracy of the prediction model parameters. In this paper, we propose sample length adaptive prediction scheme for parameter Estimation in HEVC multi-dimensional rate control. We fully consider frame difference, and set the I frame insertion interval as the length of the sample, In the calculation of the process, according sliding window, bandwidth utilization and FR of last update time to change sample length. At the same time, we focus on how to choose the initial parameters of the video initially acutely moment. We believe such a study brings us a new knowledge into the topic of multi-dimensional rate control and video transmission. The experimental results show that our model-based solutions produce better video quality compared with correlation methods.