GHOST: Grounded Human Motion Generation with Open Vocabulary Scene-and-Text Contexts.
Zoltan A. MilacskiKoichiro NiinumaRyosuke KawamuraFernando De la TorreLászló A. JeniPublished in: CoRR (2024)
Keyphrases
- human motion
- video sequences
- image sequences
- visual data
- motion capture
- human body
- keywords
- body parts
- optical motion capture
- motion sequences
- human motion tracking
- human actions
- motion recognition
- gait recognition
- three dimensional
- d scene
- spatio temporal
- dynamic scenes
- human movements
- human movement
- scene understanding
- motion synthesis
- human motion analysis
- space time
- moving objects
- camera motion
- motion segmentation
- human pose
- video scene
- object tracking
- view invariant
- viewpoint
- d objects
- facial expressions
- computer vision
- background subtraction
- video frames