CoSTA: End-to-End Comprehensive Space-Time Entanglement for Spatio-Temporal Video Grounding.
Yaoyuan LiangXiao LiangYansong TangZhao YangZiran LiJingang WangWenbo DingShao-Lun HuangPublished in: AAAI (2024)
Keyphrases
- space time
- end to end
- spatio temporal
- spatial and temporal
- video representation
- video sequences
- input video
- human actions
- temporal domain
- dynamic scenes
- multipath
- rate allocation
- motion patterns
- scalable video
- image sequences
- congestion control
- video analysis
- wireless sensor networks
- moving objects
- video clips
- human motion
- scan statistic