期刊
IEEE TRANSACTIONS ON IMAGE PROCESSING
卷 30, 期 -, 页码 6459-6472出版社
IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
DOI: 10.1109/TIP.2021.3092949
关键词
Feature extraction; Image coding; Video coding; Decoding; Convolutional codes; Transform coding; Noise reduction; Quality enhancement; compressed video; video enhancement; deep learning; attention mechanism; high efficiency video coding
The study introduces a patch-wise spatial-temporal quality enhancement network that outperforms traditional methods, achieving improved video quality in scene-changing and strong-motion videos.
Recently, many deep learning based researches are conducted to explore the potential quality improvement of compressed videos. These methods mostly utilize either the spatial or temporal information to perform frame-level video enhancement. However, they fail in combining different spatial-temporal information to adaptively utilize adjacent patches to enhance the current patch and achieve limited enhancement performance especially on scene-changing and strong-motion videos. To overcome these limitations, we propose a patch-wise spatial-temporal quality enhancement network which firstly extracts spatial and temporal features, then recalibrates and fuses the obtained spatial and temporal features. Specifically, we design a temporal and spatial-wise attention-based feature distillation structure to adaptively utilize the adjacent patches for distilling patch-wise temporal features. For adaptively enhancing different patch with spatial and temporal information, a channel and spatial-wise attention fusion block is proposed to achieve patch-wise recalibration and fusion of spatial and temporal features. Experimental results demonstrate our network achieves peak signal-to-noise ratio improvement, 0.55 - 0.69 dB compared with the compressed videos at different quantization parameters, outperforming state-of-the-art approach.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据