Perceptual Screen Content Image Quality Assessment And Compression

Shiqi Wang,Ke Gu,Kai Zeng,Zhou Wang,Weisi Lin

DOI: https://doi.org/10.1109/ICIP.2015.7351037

2015-01-01

Abstract:Compression of screen content has recently emerged as an active research topic due to the increasing demand in many applications such as wireless display and virtual desktop infrastructure. Screen content images (SCIs) exhibit different statistical properties in textual and pictorial regions, and the human visual system (HVS) also behaves differently when viewing the textual and pictorial regions in terms of the extent of visual field. Here we propose a perceptual SCI quality assessment approach that incorporates visual field adaptation and information content weighting. Furthermore, we propose a perceptual coding scheme in an attempt to optimize the HEVC Screen Content Coding encoder. Experimental results show that the proposed quality assessment method not only better predicts the perceptual quality of SCIs, but also leads to an effective way to optimize screen content coding schemes.