Time-Contrastive Networks: Self-Supervised Learning from Multi-view Observation.
Pierre SermanetCorey LynchJasmine HsuSergey LevinePublished in: CVPR Workshops (2017)
Keyphrases
- multi view
- multi view learning
- single view
- co training
- multiple cameras
- d objects
- reinforcement learning
- learning process
- supervised learning
- depth map
- range images
- multiple views
- view synthesis
- camera calibration
- semi supervised
- prior knowledge
- learning algorithm
- scene reconstruction
- free viewpoint
- learning models
- active learning
- high quality