Joint motion context and clip augmentation for spatio-temporal action detection.
Xurui MaXiang ZhangChengkun WuChuanfu XuJie LiuZhigang LuoPublished in: ICMV (2021)
Keyphrases
- spatio temporal
- action detection
- spatial and temporal
- image sequences
- space time
- moving objects
- human motion
- motion patterns
- contextual information
- human actions
- action recognition
- motion features
- motion analysis
- motion model
- video clips
- motion estimation
- key frames
- context sensitive
- three dimensional
- optical flow
- high level