AutoAlignV2: Deformable Feature Aggregation for Dynamic Multi-Modal 3D Object Detection.
Zehui ChenZhenyu LiShiquan ZhangLiangji FangQinhong JiangFeng ZhaoPublished in: CoRR (2022)
Keyphrases
- multi modal
- d objects
- object recognition
- object features
- viewpoint
- multi view
- pose estimation
- multiple views
- point distribution
- high dimensional
- cross modal
- shape descriptors
- object detection
- image features
- cad models
- multi modality
- three dimensional
- image annotation
- range data
- feature vectors
- physically plausible
- three dimensional objects
- feature set
- image analysis
- video sequences
- computer vision