Classification of Rare Building Change Using CNN with Multi-Class Focal Loss.
Keisuke NemotoRyuhei HamaguchiTomoyuki ImaizumiShuhei HikosakaPublished in: IGARSS (2018)
Keyphrases
- multi class
- multi class classifier
- support vector machine
- multi class classification
- multiclass classification
- binary and multi class
- binary classifiers
- cost sensitive
- multiple classes
- multi class boosting
- multiclass problems
- multi class svm
- multi class svms
- multi class classifiers
- class probabilities
- binary classification problems
- error correcting output codes
- feature selection
- kernel logistic regression
- single class
- binary classification
- pairwise
- machine learning
- decision trees
- classification accuracy
- binary classification tasks
- multi class problems
- probabilistic boosting tree
- support vector
- feature vectors
- protein classification
- object detection
- supervised learning
- image classification
- text classification
- svm classifier
- improve the classification accuracy
- cross validation
- semi supervised
- feature set
- training set
- text categorization
- machine learning methods
- classification algorithm