Multi-view Self-Constructing Graph Convolutional Networks with Adaptive Class Weighting Loss for Semantic Segmentation.
Qinghui LiuMichael KampffmeyerRobert JenssenArnt-Børre SalbergPublished in: CoRR (2020)
Keyphrases
- object recognition
- multi view
- semantic segmentation
- d objects
- object classes
- scene classification
- object class
- multiple views
- depth map
- computer vision
- natural images
- three dimensional
- superpixels
- weakly supervised
- conditional random fields
- semi supervised
- image features
- multiscale
- supervised learning
- graph structure
- ground truth