Disentangling Foreground and Background Motion for Enhanced Realism in Human Video Generation.
Jinlin LiuKai YuMengyang FengXiefan GuoMiaomiao CuiPublished in: CoRR (2024)
Keyphrases
- foreground and background
- foreground objects
- moving camera
- motion segmentation
- video sequences
- space time
- dynamic scenes
- background subtraction
- motion estimation
- video frames
- temporal coherence
- optical flow
- camera motion
- video data
- image sequences
- motion model
- video surveillance
- multimedia
- key frames
- mean shift
- color distribution
- moving objects
- motion analysis
- motion field
- d scene
- pixel wise
- object segmentation
- spatio temporal
- alpha matte
- bayesian framework
- human motion
- multi view
- prior knowledge
- computer vision