Weakly Supervised Semantic Segmentation of Satellite Images for Land Cover Mapping - Challenges and Opportunities.
Michael SchmittJonathan PrexlPatrick EbelLukas LiebelXiao Xiang ZhuPublished in: CoRR (2020)
Keyphrases
- satellite images
- land cover
- weakly supervised
- semantic segmentation
- multispectral
- change detection
- superpixels
- remote sensing
- conditional random fields
- topic models
- high resolution
- remote sensing images
- object class
- relation extraction
- named entities
- urban areas
- semi supervised
- multiscale
- object categories
- object detectors
- computer vision
- higher order
- object recognition
- geographic information systems
- learning algorithm
- long range
- image classification
- machine learning
- object detection
- image database
- segmentation algorithm