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: CoRR (2022)
Keyphrases
- d objects
- point cloud
- multi frame
- real world objects
- cad model
- viewpoint
- multiple views
- point correspondences
- pose estimation
- three dimensional
- point sets
- surface reconstruction
- multi view
- range data
- object recognition
- structure from motion
- image sequences
- motion estimation
- d mesh
- optical flow
- stereo camera
- motion analysis
- shape descriptors
- high resolution
- object classes
- computer vision
- machine learning
- filtering algorithm
- single image
- implicit surfaces