TransMoMo: Invariance-Driven Unsupervised Video Motion Retargeting.
Zhuoqian YangWentao ZhuWayne WuChen QianQiang ZhouBolei ZhouChen Change LoyPublished in: CoRR (2020)
Keyphrases
- temporal coherence
- space time
- motion capture
- temporal filtering
- video sequences
- spatial and temporal
- camera motion
- input video
- key frames
- data driven
- object motion
- video data
- motion estimation
- moving camera
- motion features
- temporal continuity
- low frame rate
- video frames
- video streams
- motion model
- unsupervised manner
- motion analysis
- dynamic textures
- video representation
- motion trajectories
- shot change detection
- visual data
- image sequences
- semi supervised
- motion detection
- video content
- static images
- image motion
- surveillance videos
- single frame
- successive frames
- multimedia
- visual cues
- optical flow
- periodic motion
- spatio temporal
- image frames
- motion parameters
- dynamic scenes
- motion segmentation
- video footage
- human motion
- moving objects
- video clips
- motion capture data
- temporal consistency
- video analysis
- camera movement
- image content
- motion field
- video surveillance
- video summarization
- video objects
- multi frame
- motion vectors
- video signals
- visual features
- feature points
- view invariant
- humanoid robot
- motion patterns
- layered representation
- computer vision