Geometric-aware Pretraining for Vision-centric 3D Object Detection.
Linyan HuangHuijie WangJia ZengShengchuan ZhangLiujuan CaoJunchi YanHongyang LiPublished in: CoRR (2023)
Keyphrases
- d objects
- geometric models
- multi view
- viewpoint
- multiple views
- object recognition
- topological information
- partial matching
- object detection
- three dimensional
- cad models
- pose estimation
- range data
- object parts
- computer vision
- shape descriptors
- line drawings
- cad model
- real world objects
- pose normalization
- recognition and pose estimation
- image contours
- object classes
- detection method
- physically plausible
- control points
- d mesh
- spin images
- surface properties
- geometric constraints
- object views
- vision system