Ground truth method assessment for SVM-based landscape classification.
Robin PouteauBenoît StollSébastien ChabrierPublished in: IGARSS (2010)
Keyphrases
- support vector machine svm
- classification method
- cross validation
- ground truth
- decision trees
- machine learning methods
- support vector machine
- classification accuracy
- significant improvement
- cost function
- pattern classification
- high precision
- preprocessing
- support vector
- benchmark data sets
- classification rate
- detection method
- classification procedure
- feature selection
- model selection
- high accuracy
- pattern recognition
- objective function
- test images
- ordinal data
- data sets
- ground truth data
- feature extraction
- classification scheme
- gold standard
- clustering method
- dynamic programming
- computational cost
- classification algorithm
- svm classifier
- unsupervised learning