Spatial-Temporal Adaptive Compressed Screen Content Video Quality Enhancement.
IEEE Trans. Circuits Syst. II Express Briefs(2024)
Abstract
Deep learning based methods have achieved remarkable success in compressed video quality enhancement. Multiple frames are used to extract spatial-temporal information to restore high-quality frames. However, Screen Content Videos (SCV) always have dramatic content changes, it is difficult to get accurate spatial-temporal information. To address this issue, we propose a Spatial-Temporal Adaptive (STA) method. STA can extract spatial-temporal change information in SCV, and these information can guide the adaptive fusion of single-frame and multi-frame based features, thereby not only preserving strong spatial-temporal information extraction ability but also avoiding the interference caused by content abrupt changes. Specifically, we introduce a dual-branch structure for single-frame and multi-frame feature extraction in parallel. An adaptive feature fusion module is also introduced to fuse the features from two branches with the guidance of spatial-temporal change information. Experimental results demonstrate that our method outperforms previous methods in compressed SCV quality enhancement.
MoreTranslated text
Key words
Screen content video,compressed video quality enhancement,dual-branch network,feature fusion
AI Read Science
Must-Reading Tree
Example
Generate MRT to find the research sequence of this paper
Chat Paper
Summary is being generated by the instructions you defined