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