BEV-CFKT: A LiDAR-camera cross-modality-interaction fusion and knowledge transfer framework with transformer for BEV 3D object detection.
Ming WeiJiachen LiHongyi KangYijie HuangJun-Guo LuPublished in: Neurocomputing (2024)
Keyphrases
- d objects
- knowledge transfer
- viewpoint
- multi view
- object recognition
- cross domain
- transfer learning
- three dimensional
- shape descriptors
- multiple views
- pose estimation
- image classification
- cad models
- range data
- machine learning algorithms
- social networks
- object detection
- feature space
- decision trees
- decision making
- feature selection