CFI2P: Coarse-to-Fine Cross-Modal Correspondence Learning for Image-to-Point Cloud Registration.
Gongxin YaoYixin XuanYiwei ChenYu PanPublished in: CoRR (2023)
Keyphrases
- point cloud
- coarse to fine
- multiscale
- cross modal
- image matching
- image retrieval
- image registration
- matching scheme
- multiresolution
- learning algorithm
- reinforcement learning
- image content
- image analysis
- real world objects
- image data
- point sets
- multi modal
- input image
- image regions
- image classification
- supervised learning
- image segmentation
- visual recognition
- laser scanner
- image representation
- image database
- image features
- high resolution
- similarity measure