D-Align: Dual Query Co-attention Network for 3D Object Detection Based on Multi-frame Point Cloud Sequence.
Junhyung LeeJunho KohYoungwoo LeeJun Won ChoiPublished in: ICRA (2023)
Keyphrases
- d objects
- point cloud
- multi frame
- real world objects
- viewpoint
- object recognition
- multiple views
- multi view
- cad model
- point correspondences
- range data
- point sets
- surface reconstruction
- three dimensional
- pose estimation
- shape descriptors
- motion estimation
- super resolution
- object surface
- structure from motion
- multiscale
- computationally efficient
- object classes
- d mesh
- stereo camera
- image sequences
- machine learning
- high resolution
- optic flow
- implicit surfaces