Optimal Decision Fusion for a Face Verification System

Fusion is a popular practice to increase the reliability of the biometric verification. In this paper, optimal fusion at decision level by AND rule and OR rule is investigated. Both a theoretical analysis and the experimental results are given. Comparisons are presented between fusion at decision level and fusion at matching score level. For our face verification system, decision fusion proves to be a simple, practical, and effective approach, which significantly improves the performance of the original classifier.

• Type: Conference
• Year: 2007
• Where: ICB
• Authors: Qian Tao, Raymond N. J. Veldhuis
• Keywords: Biometrics, Decision Level, Face Verification System, ICB 2007, Optimal Fusion