Optimization Methods for Identity Verification System Using Biometric Features

The paper will present an approach to solve the real-world biometric identification problems, with an objective to circumvent the curse of dimensionality in biometric identification system through very large database size. The main objective of the proposed approach is obtaining of effective in reducing the complexity and hence improving the performance of user identification. This will be achieved through the following steps: Firstly, reduction of features extraction, we will use feature selection algorithm based on several objective optimization techniques. Secondly, acceleration of classification, we will use recent method to determine clusters of users having significant similarities and then apply Support Vector Machine (SVM) classifiers on each cluster of users independently. Hence reducing the dimensionality of the dataset for training SVMs and improving the performance of user identification. The experimental results of the proposed technique are presented and compared.