STDiff: Spatio-temporal Diffusion for Continuous Stochastic Video Prediction.
Xi YeGuillaume-Alexandre BilodeauPublished in: CoRR (2023)
Keyphrases
- spatio temporal
- spatial and temporal
- space time
- spatial temporal
- video representation
- human actions
- prediction accuracy
- video sequences
- video data
- video frames
- spatio temporally
- video streams
- temporal domain
- multimedia
- prediction error
- prediction model
- video content
- anisotropic diffusion
- video surveillance
- real time
- prediction algorithm
- moving objects
- video retrieval
- monte carlo
- image sequences
- continuous domains
- video analysis
- temporal segmentation
- spatial and temporal relationships
- motion trajectories
- multiscale
- diffusion process
- motion patterns
- scalable video coding
- key frames
- temporal consistency
- multimedia data
- activity recognition
- action recognition
- spatio temporal queries
- computer vision