Incorporating Depth into both CNN and CRF for Indoor Semantic Segmentation.
Jindong JiangZhijun ZhangYongqian HuangLunan ZhengPublished in: CoRR (2017)
Keyphrases
- semantic segmentation
- conditional random fields
- street scenes
- superpixels
- scene classification
- label transfer
- weakly supervised
- object categories
- long range
- depth map
- object classes
- pascal voc
- information extraction
- object class
- higher order
- graphical models
- generative model
- object recognition
- pairwise
- viewpoint
- image set
- markov random field
- probabilistic model
- object detection
- image classification
- image understanding
- computer vision