Adaptive background skipping algorithm for region-of-interest scalable video coding

Abstract

In this paper, we propose an adaptive region of interest (ROI) algorithm for scalable video coding. ROIs are encoded in every temporal layer, while background (non-ROI) is skipped selectively in enhancement layers. The background skipping decision is based on an adaptive estimation of motion activity. And the saving bits could be used to enhance the ROI spatial quality. The overall perceptual quality can be significantly improved since human vision tends to pay more attention to ROIs. The experimental results indicate the proposed scheme has better visual quality compared with conventional temporal scalable scheme.