Article: Fusing multiple matcher|s outputs for secure human identification Journal: International Journal of Biometrics (IJBM) 2009 Vol.1 No.3 pp.329 - 348 Abstract: Multimodal biometrics is an emerging area of research that aims at increasing the reliability of biometric systems through utilising more than one biometric in decision-making process. An effective fusion scheme plays a key role in combining the information presented by the multiple domain experts. Such information can be integrated at several distinct levels, such as sensor level, feature level, match score level, rank level and decision level. This paper describes the combination process of different monomodal expert through rank and decision fusion methods using iris, ear and face biometrics for secure human authentication. For rank-level fusion, plurality voting, Borda count and logistic regression approaches are employed and compared, and for decision-level fusion, AND/OR, majority voting, weighted majority voting and behavioural knowledge space approaches have been implemented and tested. The key contribution of the paper is in comparison of the recognition performance of the developed multimodal system for all of the above approaches. The results indicate that fusing individual modalities improve the overall performance of the biometric system and the logistic regression rank-level fusion results in the highest recognition performance even in the presence of low-quality data. Inderscience Publishers - linking academia, business and industry through research

Title: Fusing multiple matcher's outputs for secure human identification

Authors: Marina L. Gavrilova, Md. Maruf Monwar

Addresses: Department of Computer Science, University of Calgary, 2500 University Drive NW, Calgary, AB T2N 1N4, Canada. ' Department of Computer Science, University of Calgary, 2500 University Drive NW, Calgary, AB T2N 1N4, Canada

Abstract: Multimodal biometrics is an emerging area of research that aims at increasing the reliability of biometric systems through utilising more than one biometric in decision-making process. An effective fusion scheme plays a key role in combining the information presented by the multiple domain experts. Such information can be integrated at several distinct levels, such as sensor level, feature level, match score level, rank level and decision level. This paper describes the combination process of different monomodal expert through rank and decision fusion methods using iris, ear and face biometrics for secure human authentication. For rank-level fusion, plurality voting, Borda count and logistic regression approaches are employed and compared, and for decision-level fusion, AND/OR, majority voting, weighted majority voting and behavioural knowledge space approaches have been implemented and tested. The key contribution of the paper is in comparison of the recognition performance of the developed multimodal system for all of the above approaches. The results indicate that fusing individual modalities improve the overall performance of the biometric system and the logistic regression rank-level fusion results in the highest recognition performance even in the presence of low-quality data.

Keywords: multimodal biometrics; pattern recognition; biometric fusion; logistic regression; behavioural knowledge space; weighted majority voting; multiple outputs; matcher outputs; human identification; iris recognition; ear recognition; face recognition; for security; human authentication.

DOI: 10.1504/IJBM.2009.024277

International Journal of Biometrics, 2009 Vol.1 No.3, pp.329 - 348

Published online: 30 Mar 2009 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article

Title: Fusing multiple matcher's outputs for secure human identification

Keep up-to-date