Enhancing Single-Frame Supervision for Better Temporal Action Localization.
Changjian ChenJiashu ChenWeikai YangHaoze WangJohannes KnittelXibin ZhaoSteffen KochThomas ErtlShixia LiuPublished in: CoRR (2023)
Keyphrases
- single frame
- temporal redundancy
- multi frame
- super resolution
- video sequences
- multiple frames
- optical flow
- low resolution
- high resolution video
- temporal information
- video frames
- image sequences
- spatio temporal
- temporal coherence
- human actions
- multiscale
- video coding
- surveillance videos
- high resolution
- object recognition