Automatic Selection of the Training Set for Semi-supervised Land Classification and Segmentation of Satellite Images.
Olga RajadellPedro García-SevillaPublished in: ICPRAM (1) (2012)
Keyphrases
- satellite images
- automatic selection
- training set
- landsat tm
- semi supervised
- supervised learning
- classification accuracy
- multispectral
- land cover
- remote sensing
- classification algorithm
- training samples
- remotely sensed
- change detection
- unlabeled data
- active learning
- cross validation
- semi supervised learning
- supervised classification
- semi supervised classification
- decision trees
- class labels
- multi spectral images
- machine learning
- data sets
- training data
- pattern recognition
- feature space
- level set
- segmentation algorithm
- decision boundary
- support vector machine
- high resolution
- hyperspectral
- high resolution satellite images
- support vector
- labeled data
- unsupervised learning
- svm classifier
- support vector machine svm
- test set
- training examples
- unsupervised clustering
- text classification
- hyperspectral data
- image processing
- image segmentation
- image analysis
- test images
- urban areas
- hyperspectral images
- feature vectors
- co training
- remote sensing images
- satellite imagery
- nearest neighbor
- digital elevation models
- feature selection
- fully supervised
- computer vision
- medical imaging