JPV-Net: Joint Point-Voxel Representations for Accurate 3D Object Detection.
Nan SongTianyuan JiangJian YaoPublished in: AAAI (2022)
Keyphrases
- d objects
- three dimensional
- viewpoint
- object recognition
- control points
- surface points
- multi view
- pose estimation
- multiple views
- geometric models
- shape descriptors
- minimum bounding
- range data
- d mesh
- object representation
- cad models
- object detection
- image contours
- line drawings
- object classes
- physically plausible
- viewing direction
- detection method
- partial matching
- point distribution
- pose normalization
- multi views
- geometric invariants
- matching algorithm
- face recognition