Pointly-supervised 3D Scene Parsing with Viewpoint Bottleneck.
Liyi LuoBeiwen TianHao ZhaoGuyue ZhouPublished in: CoRR (2021)
Keyphrases
- d scene
- viewpoint
- camera viewpoint
- single image
- depth map
- scene understanding
- planar surfaces
- scene structure
- camera parameters
- semi supervised
- learning algorithm
- image based rendering
- flow estimation
- multiple views
- photorealistic
- feature selection
- optical flow
- view synthesis
- view dependent
- object class
- camera positions
- indoor scenes
- coordinate frame
- machine learning
- data sets
- d objects
- active learning
- pairwise
- scene flow