F. Alsaade, A. Rahmoun, and A. Ech-Cherif(Saudi Arabia)
Multimodal biometrics fusion, Unconstrained Cohort Normalization (UCN), Brute Force Search (BFS), Genetic Algorithms (GA), Neural Networks (NN), MinMax Normalization.
In this paper, we introduce hybrid intelligent techniques such as neural networks and genetic algorithms to enhance performance of multimodal biometric fusion. Several techniques have been intensively investigated in the last few years towards fusion of multimodal biometrics. We propose to introduce neural networks for their "learning" capabilities to a better verification for a decision level, whereas genetic algorithms are used to generate optimal weights of multimodal biometrics fusion. It is clearly shown that combining these techniques along with conventional fusion schemes such as Unconstrained Cohort Normalization lead to better performance at a decision level. It is shown in this paper that by deploying such technique at the decision level, the system error rate can be reduced considerably. The experimental investigations involve the recognition mode of verification in mixed quality data conditions. The paper presents the motivation, and the potential advantages of the proposed approach and details about the experimental study.
Important Links:
Go Back
