A Multilevel-Guided Curriculum Domain Adaptation Approach to Semantic Segmentation for High-Resolution Remote Sensing Images.
Zhihao XiXiangyu HeYu MengAnzhi YueJingbo ChenYupeng DengJiansheng ChenPublished in: IEEE Trans. Geosci. Remote. Sens. (2023)
Keyphrases
- domain adaptation
- semantic segmentation
- conditional random fields
- superpixels
- cross domain
- multiple sources
- labeled data
- semi supervised
- scene classification
- transfer learning
- sentiment classification
- semi supervised learning
- object categories
- test data
- information extraction
- training data
- co training
- target domain
- unlabeled data
- object classes
- high resolution
- prior knowledge
- machine learning
- natural images
- feature extraction
- bayesian networks
- object recognition
- input image
- general purpose
- image classification
- data sets