LaMD: Latent Motion Diffusion for Video Generation.
Yaosi HuZhenzhong ChenChong LuoPublished in: CoRR (2023)
Keyphrases
- space time
- spatial and temporal
- key frames
- object motion
- video sequences
- temporal filtering
- moving camera
- motion features
- input video
- visual cues
- successive frames
- low frame rate
- visual data
- video frames
- video representation
- image sequences
- dynamic scenes
- shot change detection
- human motion
- dynamic textures
- video content
- optical flow
- video summarization
- real time
- temporal consistency
- single frame
- motion analysis
- video streams
- temporal coherence
- motion patterns
- video data
- multimedia
- motion trajectories
- static images
- video footage
- spatio temporal
- video analysis
- human actions
- reference frame
- moving objects
- periodic motion
- temporal continuity
- motion estimation
- camera motion
- video signals
- video retrieval
- diffusion process
- video clips
- wide area motion imagery
- surveillance videos
- video processing
- visual features
- motion capture data
- temporal domain
- motion model
- motion field
- anisotropic diffusion
- video scene
- motion segmentation
- humanoid robot
- action recognition
- image frames
- video shots