An Adaptive Hierarchical QP Setting for Screen Content Coding

Jiahao Li,Bin Li,Jizheng Xu,Ruiqin Xiong

DOI: https://doi.org/10.1109/vcip.2015.7457812

2015-01-01

Abstract:Screen content refers to computer generated content like text, graphics, and animations. In such video, many regions may remain static for a long period after a sudden change. Traditional hierarchical Quantization Parameter (QP) setting may not be able to handle these regions efficiently because the encoder probably needs to refine the quality of these static regions multiple times. It will cost more bits while the quality of the static regions may reach the expected degree which the flat QP setting is able to achieve. This paper proposes using different QP settings for different regions in a picture. Region classification algorithms are developed to determine whether a flat or hierarchical QP setting is used. Experimental results demonstrate that the proposed scheme can achieve an average bitrate reduction of 3.1%, and up to 8.1% bitrate reduction for IBBB coding. The proposed method improves coding efficiency without increasing encoding complexity.