An active learning heuristic using spectral and spatial information for MRF-based classification.
Bo HuGabriele MoserSebastiano B. SerpicoPeijun LiPublished in: IGARSS (2015)
Keyphrases
- spatial information
- active learning
- spatial distribution
- spatial relationships
- spatial resolution
- supervised learning
- selective sampling
- spatial relations
- decision trees
- feature extraction
- classification accuracy
- temporal information
- spatial features
- local binary pattern
- intensity values
- color histogram
- markov random field
- region connection calculus
- machine learning
- support vector machine
- feature selection
- image classification
- training set
- feature space
- training data
- feature set
- relevance feedback
- hyperspectral
- learning process
- feature vectors
- rare class