PG-RCNN: Semantic Surface Point Generation for 3D Object Detection.
Inyong KooInyoung LeeSe-Ho KimHeeseon KimWoo-Jin JeonChangick KimPublished in: ICCV (2023)
Keyphrases
- d objects
- surface points
- viewpoint
- range data
- object recognition
- control points
- three dimensional
- viewing direction
- multi view
- multiple views
- shape descriptors
- pose estimation
- geometric models
- d mesh
- surface features
- object detection
- minimum bounding
- cad models
- image contours
- surface patches
- object views
- computer vision
- physically plausible
- pose normalization
- texture mapping
- cross section
- line drawings
- semantic information
- object features
- real world objects
- three dimensional objects
- topological information
- curved objects
- object classes
- computer graphics