A Bayesian-inspired, deep learning-based, semi-supervised domain adaptation technique for land cover mapping.
Benjamin LucasCharlotte PelletierDaniel F. SchmidtGeoffrey I. WebbFrançois PetitjeanPublished in: Mach. Learn. (2023)
Keyphrases
- training data
- domain adaptation
- deep learning
- land cover
- semi supervised
- labeled data
- unsupervised learning
- supervised classification
- semi supervised learning
- unlabeled data
- supervised learning
- weakly supervised
- remote sensing
- multispectral
- change detection
- satellite images
- active learning
- co training
- machine learning
- training examples
- learning algorithm
- geographic information systems
- bayesian networks
- transfer learning
- training set
- pairwise
- data sets
- multi view
- generative model
- maximum likelihood
- multiple sources
- video sequences
- cross domain
- co occurrence
- multiscale
- feature selection
- data streams