Render In-between: Motion Guided Video Synthesis for Action Interpolation.
Hsuan-I HoXu ChenJie SongOtmar HilligesPublished in: BMVC (2021)
Keyphrases
- recognizing human actions
- human actions
- space time
- action recognition
- recognition of human actions
- human motion
- key frames
- temporal domain
- motion features
- object motion
- action classification
- temporal filtering
- input video
- spatial and temporal
- video sequences
- body motions
- multimedia
- visual cues
- video content
- motion history images
- motion synthesis
- motion analysis
- video frames
- video data
- successive frames
- low frame rate
- spatio temporal
- visual data
- shot change detection
- static images
- image frames
- surveillance videos
- image sequences
- moving camera
- camera movement
- motion model
- dynamic textures
- visual features
- action detection
- temporal consistency
- dynamic scenes
- motion estimation
- temporal continuity
- layered representation
- motion trajectories
- video objects
- moving objects
- video streams
- motion field
- reference frame
- displacement estimation
- motion capture data
- video surveillance
- video retrieval
- temporal coherence
- mechanical devices
- video signals
- image interpolation
- optical flow
- video shots
- human body
- camera motion
- single frame
- video analysis
- atomic actions
- video footage
- motion patterns
- video summarization