V-DETR: DETR with Vertex Relative Position Encoding for 3D Object Detection.
Yichao ShenZigang GengYuhui YuanYutong LinZe LiuChunyu WangHan HuNanning ZhengBaining GuoPublished in: CoRR (2023)
Keyphrases
- d objects
- relative position
- viewpoint
- multi view
- object recognition
- pose estimation
- multiple views
- shape descriptors
- three dimensional
- range data
- line drawings
- d mesh
- cad models
- geometric properties
- physically plausible
- object detection
- object classes
- video sequences
- image contours
- topological information
- object views
- pose normalization
- minimum bounding
- machine learning
- feature extraction
- geometric models
- image segmentation