Motion-Augmented Self-Training for Video Recognition at Smaller Scale.
Kirill GavrilyukMihir JainIlia KarmanovCees G. M. SnoekPublished in: ICCV (2021)
Keyphrases
- space time
- static images
- video sequences
- motion features
- key frames
- video data
- spatial and temporal
- human activities
- input video
- recognition rate
- object motion
- motion estimation
- temporal filtering
- moving camera
- image sequences
- motion analysis
- shot change detection
- successive frames
- human motion
- visual data
- action recognition
- temporal continuity
- video frames
- temporal consistency
- video clips
- video objects
- visual cues
- temporal segmentation
- motion trajectories
- human gait
- low frame rate
- visual perception
- surveillance videos
- gait recognition
- dynamic scenes
- recognition algorithm
- video streams
- activity recognition
- object recognition
- multimedia
- motion history images
- motion patterns
- human actions
- feature extraction
- motion recognition
- camera motion
- motion model
- hand gestures
- video signals
- motion parameters
- motion capture data
- video shots
- optical flow
- video content
- temporal coherence
- video summarization
- video analysis
- dynamic textures
- computer vision
- scale space
- camera movement
- semi supervised
- co training
- spatio temporal
- wide area motion imagery