Semantic Segmentation of Sparsely Annotated 3D Point Clouds by Pseudo-Labelling.
Katie XuYasuhiro YaoKazuhiko MurasakiShingo AndoAtsushi SagataPublished in: 3DV (2019)
Keyphrases
- semantic segmentation
- street scenes
- image set
- conditional random fields
- superpixels
- label transfer
- scene classification
- weakly supervised
- object categories
- depth map
- manually annotated
- object classes
- object class
- object recognition
- image understanding
- computer vision
- pixel level
- natural images
- annotated images
- multiscale
- segmentation method
- generative model
- object detection
- hidden markov models