MoCoGAN: Decomposing Motion and Content for Video Generation.
Sergey TulyakovMing-Yu LiuXiaodong YangJan KautzPublished in: CoRR (2017)
Keyphrases
- video footage
- space time
- multimedia
- spatial and temporal
- temporal filtering
- object motion
- input video
- online video
- video sequences
- key frames
- video content
- video frames
- dynamic scenes
- generation process
- multimedia data
- video representation
- video data
- motion estimation
- moving camera
- low frame rate
- temporal consistency
- visual cues
- successive frames
- shot change detection
- dynamic textures
- video streams
- motion detection
- temporal continuity
- static images
- motion analysis
- video segments
- video material
- temporal coherence
- video clips
- surveillance videos
- video analysis
- human motion
- metadata
- motion patterns
- image sequences
- moving objects
- camera motion
- temporal domain
- visual data
- video retrieval
- video scene
- single frame
- motion features
- video database
- user generated
- multimedia content
- reference frame
- layered representation
- motion capture
- camera movement
- spatio temporal
- motion trajectories
- lecture videos
- video objects
- human activities
- motion model
- motion segmentation
- image frames
- video shots
- image motion
- global motion
- real time