Cross-Identity Motion Transfer for Arbitrary Objects Through Pose-Attentive Video Reassembling.
Subin JeonSeonghyeon NamSeoung Wug OhSeon Joo KimPublished in: ECCV (24) (2020)
Keyphrases
- object motion
- rigid objects
- dynamic scenes
- static images
- visual data
- moving objects
- space time
- motion trajectories
- video objects
- video sequences
- d objects
- visual input
- articulated objects
- object detection and tracking
- layered representation
- temporal filtering
- spatial and temporal
- temporal continuity
- stationary camera
- video data
- focus of attention
- object pose
- motion features
- image sequences
- wide area motion imagery
- target object
- key frames
- motion estimation
- pose estimation
- object shape
- motion model
- motion segmentation
- video content
- visual attention
- dynamic textures
- motion analysis
- human motion
- partial occlusion
- human pose
- visual cues
- moving camera
- optical flow
- appearance variations
- video surveillance
- camera motion
- visual motion
- video images
- bounding box
- long video sequences
- motion parameters
- multi view
- spatio temporal
- video frames
- video scene
- position and orientation
- motion patterns
- object appearance
- video streams
- humanoid robot
- object segmentation