AutoAlign: Pixel-Instance Feature Aggregation for Multi-Modal 3D Object Detection.
Zehui ChenZhenyu LiShiquan ZhangLiangji FangQinhong JiangFeng ZhaoBolei ZhouHang ZhaoPublished in: CoRR (2022)
Keyphrases
- multi modal
- d objects
- object features
- object recognition
- pose estimation
- viewpoint
- multi view
- point distribution
- three dimensional
- shape descriptors
- d mesh
- cad models
- multi modality
- multiple views
- high dimensional
- image features
- physically plausible
- object classes
- computer vision
- cross modal
- image contours
- pose normalization
- range data
- object detection
- uni modal
- integral imaging
- single modality
- visual features
- feature vectors
- image sequences
- image processing