DCSN: Deformable Convolutional Semantic Segmentation Neural Network for Non-Rigid Scenes.
Bor-Sheng HuangChih-Chung HsuWo-Ting LiaoHan-Yi KaoXian-Yun WangPublished in: ICASSP (2022)
Keyphrases
- semantic segmentation
- street scenes
- superpixels
- conditional random fields
- scene classification
- label transfer
- weakly supervised
- object classes
- deformable models
- natural scenes
- object categories
- multiple objects
- pascal voc
- object class
- feature points
- segmentation algorithm
- d objects
- multiscale
- image set
- image understanding
- segmentation method
- visual features
- graphical models
- higher order
- active learning
- object recognition