Biometric based personal authentication using fuzzy binary decision tree

This paper proposes a fuzzy binary decision tree (FBDT) algorithm for biometric based personal authentication. The proposed FBDT algorithm utilizes fuzzified matching scores to construct a decision tree which aimed at decision making in two classes: genuine and imposter. The membership values of the matching scores can be automatically computed and the tree-nodes can be learned with the enrolment of the new users in the database. The performance of the FBDT is examined on publicly available databases and found superior than its crisp equivalent.