DynaVol: Unsupervised Learning for Dynamic Scenes through Object-Centric Voxelization.
Yanpeng ZhaoSiyu GaoYunbo WangXiaokang YangPublished in: ICLR (2024)
Keyphrases
- dynamic scenes
- unsupervised learning
- moving objects
- complex dynamic scenes
- rigid objects
- multi view
- static scenes
- video sequences
- motion segmentation
- space time
- d objects
- background subtraction
- motion cues
- supervised learning
- object tracking
- moving camera
- spatio temporal
- frame rate
- machine learning
- feature selection
- semi supervised
- multiple views
- optical flow
- foreground objects
- face recognition
- computer vision