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