GVA: Reconstructing Vivid 3D Gaussian Avatars from Monocular Videos.
Xinqi LiuChenming WuJialun LiuXing LiuJinbo WuChen ZhaoHaocheng FengErrui DingJingdong WangPublished in: CoRR (2024)
Keyphrases
- image sequences
- video sequences
- view invariant
- virtual environment
- virtual world
- pose estimation
- virtual reality
- video frames
- video content
- maximum likelihood
- pose tracking
- model based tracking
- stereo camera
- video database
- video analysis
- optical flow
- gaussian distribution
- human activities
- moving camera
- gaussian noise
- monocular images
- space time
- independently moving objects
- video data
- dynamic scenes
- gaussian mixture model
- action recognition
- user generated
- video segments
- depth cues