Improving the performance of lightweight CNNs for binary classification using quadratic mutual information regularization.
Maria TzelepiAnastasios TefasPublished in: Pattern Recognit. (2020)
Keyphrases
- lightweight
- binary classification
- mutual information
- multi class
- support vector
- learning problems
- cost sensitive
- support vector machine
- multi label
- multi class classification
- feature selection
- image registration
- binary classifiers
- convex loss functions
- class imbalance
- prediction accuracy
- pairwise
- generalization error
- similarity measure
- wireless sensor networks
- machine learning algorithms
- image classification
- learning tasks
- ensemble methods
- machine learning
- data mining
- neural network
- reproducing kernel hilbert space
- data sets