A Robust Hierarchical Qp Setting for Screen Content Coding

Tong Tang; Ling Li; Donglin Wang

ICASSP 2018accepted0 citations

A Robust Hierarchical Qp Setting for Screen Content Coding

Abstract

Unlike conventional camera-captured videos (CCV), the screen content videos (SCV) are generated by computer, like text, animation, or graphics. The video contents are discontinuous, as one abrupt frame is often followed by many static frames. Therefore, the traditional hierarchical quantization parameter (QP) setting for CCV may not be suitable for SCV. In this paper, a robust hierarchical QP setting method is proposed. Concretely, a region recognition method is designed to identify the influential regions, whose quality has tremendous influence on subsequent frames. These influential regions employ a smaller QP to improve the RD (Rate Distortion) performance of encoded video. Experimental results show that the proposed method averagely reduces 1.7% bitrate, and improves RD performance for all the SCV, which outperforms other encoding methods.

BibTeX

@inproceedings{icassp2018_arobusthierarchi,
  title = {A Robust Hierarchical Qp Setting for Screen Content Coding},
  author = {Tong Tang and Ling Li and Donglin Wang},
  booktitle = {ICASSP 2018},
  year = {2018}
}