PointCSE: Context-sensitive encoders for efficient 3D object detection from point cloud.
Kuoliang WuGuodong XuZili LiuHaifeng LiuDeng CaiXiaofei HePublished in: Int. J. Mach. Learn. Cybern. (2022)
Keyphrases
- context sensitive
- d objects
- point cloud
- real world objects
- pose estimation
- cad model
- object recognition
- viewpoint
- range data
- three dimensional
- surface reconstruction
- multi view
- structure from motion
- multiple views
- d mesh
- point sets
- language model
- shape descriptors
- laser scanner
- computer vision
- active learning
- object surface
- machine learning
- hough transform
- supervised learning
- single view
- stereo camera
- moving objects