Noise-Resistant Deep Learning for Object Classification in Three-Dimensional Point Clouds Using a Point Pair Descriptor.
Dmytro BobkovSili ChenRuiqing JianMuhammad Zafar IqbalEckehard G. SteinbachPublished in: IEEE Robotics Autom. Lett. (2018)
Keyphrases
- point cloud
- object classification
- deep learning
- noise resistant
- feature hierarchies
- three dimensional
- object recognition
- unsupervised learning
- structure from motion
- machine learning
- d objects
- point sets
- object categories
- weakly supervised
- object detection
- pedestrian detection
- multi view
- keypoints
- viewpoint
- mental models
- semi supervised
- action recognition
- human body
- feature extraction
- image sequences