Unsupervised Domain Adaptation for Monocular 3D Object Detection via Self-training.
Zhenyu LiZehui ChenAng LiLiangji FangQinhong JiangXianming LiuJunjun JiangPublished in: ECCV (9) (2022)
Keyphrases
- d objects
- pose estimation
- object recognition
- multi view
- cad models
- multiple views
- viewpoint
- three dimensional
- shape descriptors
- image contours
- image sequences
- line drawings
- range data
- physically plausible
- surface points
- object classes
- machine learning
- object detection
- training set
- shape matching
- complex background
- surface patches
- detection method
- geometric invariants
- pose normalization
- computer vision