PatchContrast: Self-Supervised Pre-training for 3D Object Detection.
Oren ShroutOri NizanYizhak Ben-ShabatAyellet TalPublished in: CoRR (2023)
Keyphrases
- d objects
- multi view
- viewpoint
- pose estimation
- object recognition
- range data
- multiple views
- three dimensional
- shape descriptors
- cad models
- line drawings
- object model
- object classes
- pose normalization
- d mesh
- training examples
- machine learning
- control points
- three dimensional objects
- object contours
- object detection
- target detection
- object representation
- real world objects
- cad model
- object views
- integral imaging
- training set
- physically plausible