On using error bounds to optimize cost-sensitive multimodal biometric authentication.
Norman PohJosef KittlerPublished in: ICPR (2008)
Keyphrases
- error bounds
- cost sensitive
- biometric authentication
- directly optimize
- misclassification costs
- cost sensitive learning
- multi class
- biometric authentication systems
- biometric systems
- cost sensitive classification
- theoretical analysis
- worst case
- multimodal biometrics
- active learning
- class distribution
- user authentication
- naive bayes
- class imbalance
- support vector machine
- upper bound
- multi label
- machine learning
- cost sensitive boosting
- feature selection
- data mining