0%

(CVPR 2017) Real-Time Video Super-Resolution with Spatio-Temporal Networks and Motion Compensation

Keyword [VESPCN] [Pixel Shuffle] [Optical Flow] [FlowNet]

Caballero J, Ledig C, Aitken A, et al. Real-time video super-resolution with spatio-temporal networks and motion compensation[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 2017: 4778-4787.



1. Overview


基于ESPCN,论文

  • 扩展到spatial-temporal
  • 使用STN进行motion compensation

1.1. 模型





1.2. Loss Function




1.3. Fusion方案



1.4. 数据集

  • CDVL. 115 uncompressed full HD videos

1.5. 评价指标

  • MOVIE. Motion-based Video Integrity Evaluation
    The MOVIE index was designed as a metric measuring video quality that correlates with human perception and incorporates a notion of temporal consistency.


1.6. 实验结果

  • L. Layer number
  • E5. Early fusion with 5 frame
  • MC. Motion compensation