Fuzzy matching engine for non-textual authentication: A case study

The familiar paradigm of authentication by means of a textual username and password has been questioned for some years but it is only relatively recently, with the advent of technological advances in such areas as biometric and gesture recognition that truly alternative approaches to authentication have become feasible. The primary difference between nontextual methods and the classical approach to authentication is that the latter respond only to an exactly matching password while the former must respond to gestures that do not exactly match, but are sufficiently close to the stored prototype. Under this new constraint, the use of fuzzy logic appears to be an appropriate approach to authenticating a user. This paper describes a real application for logging into a secure system. A fuzzy matching algorithm was developed that matches gestures entered at the login prompt against previously registered gestures held in a database.