RVT: Robotic View Transformer for 3D Object Manipulation.
Ankit GoyalJie XuYijie GuoValts BlukisYu-Wei ChaoDieter FoxPublished in: CoRR (2023)
Keyphrases
- d objects
- multiple views
- object views
- viewpoint
- manipulation tasks
- multi view
- object recognition
- pose estimation
- three dimensional
- shape descriptors
- cad models
- image contours
- range data
- d mesh
- line drawings
- geometric models
- single view
- viewing direction
- topological information
- control points
- surface points
- object model
- image features
- geometric invariants
- pose normalization
- physically plausible