An Adaptive Fusion Framework for Fault-Tolerant Multibiometrics

Chindaro, Samuel and Zhou, Ziheng and Ng, Ming Wah R. and Deravi, Farzin (2009) An Adaptive Fusion Framework for Fault-Tolerant Multibiometrics. In: International Conference on Information Security and Digital Forensics 2009, 7-8 September 2009, City University, London. (The full text of this publication is not available from this repository)

The full text of this publication is not available from this repository. (Contact us about this Publication)

Abstract

The use of multiple biometrics will work with greater efficiency if all the systems are capable of acquiring biometrics of adequate quality and processing them successfully. However if one or more of the biometrics fails, then the system has to rely on fewer or one biometric. If the individual biometrics are set to use low thresholds, the system maybe vulnerable to falsely accepting impostors. The motivation behind the proposed method is to provide an adaptive fusion platform where the software system can identify failures in certain algorithms and if necessary adapt the current rule to ignore these algorithms and adjust operating points accordingly. Results from experiments carried out on a multi-algorithmic and multi-biometric 3D and 2D database are presented to show that adopting such a system will result in an improvement in efficiency and verification rate.

Item Type: Conference or workshop item (Paper)
Subjects: T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK7800 Electronics (see also: telecommunications) > TK7880 Applications of electronics (inc industrial & domestic) > TK7882.B56 Biometrics
Divisions: Faculties > Science Technology and Medical Studies > School of Engineering and Digital Arts > Image and Information Engineering
Depositing User: J. Harries
Date Deposited: 11 Jan 2010 15:16
Last Modified: 15 Jul 2014 13:58
Resource URI: http://kar.kent.ac.uk/id/eprint/23511 (The current URI for this page, for reference purposes)
  • Depositors only (login required):