How Low Can We Go? Pixel Annotation for Semantic Segmentation.
Daniel KigliAriel ShamirShai AvidanPublished in: CoRR (2022)
Keyphrases
- semantic segmentation
- label transfer
- street scenes
- superpixels
- pixel level
- conditional random fields
- weakly supervised
- scene classification
- object categories
- pascal voc
- image understanding
- object detection
- input image
- object classes
- segmentation algorithm
- object class
- generative model
- image retrieval
- image set
- learning algorithm
- pixel wise
- ground truth
- viewpoint
- object recognition