MEViT: Motion Enhanced Video Transformer for Video Classification.
Li LiLiansheng ZhuangPublished in: MMM (2) (2022)
Keyphrases
- video classification
- dynamic textures
- video content
- motion trajectories
- video shots
- video clips
- semantic video
- key frames
- web videos
- space time
- image sequences
- motion estimation
- spatial and temporal
- video indexing
- video sequences
- motion patterns
- motion model
- camera motion
- optical flow
- motion analysis
- human motion
- video frames
- video retrieval
- video streams
- moving objects
- video database
- video search
- feature points
- dynamic scenes
- visual data
- video segments