MixSup: Mixed-grained Supervision for Label-efficient LiDAR-based 3D Object Detection.
Yuxue YangLue FanZhaoxiang ZhangPublished in: ICLR (2024)
Keyphrases
- d objects
- multi view
- object recognition
- pose estimation
- viewpoint
- physically plausible
- three dimensional
- multiple views
- detection method
- object detection
- d mesh
- shape descriptors
- lidar data
- range data
- cad models
- object features
- object classes
- pose normalization
- integral imaging
- control points
- line drawings
- high resolution
- active learning