Biometric Access Control uses quantative anatomical, physiological or behavioral traits to automatically and accurately verify the identity of an individual. The goal is to control or restrict acces to appropriate individuals. Accuracy and reliabilty requirements are high, and today, such systems play an increasingly important role in authentication and authorization of individuals. Face recognition has long been regarded as anideal biometric resource, but the complexity of analysis and comparison of facial characteristics has limited successful application of 2D and other current techniques in this domain. Typical 2D applications currently in use have high error rates and are hampered by variations such as lighting and orientation. This project will develop and apply true 3D spatial analysis tools and techniques to acurately recognize and compare distinct 3D facial features.. In addition, we will create, maintain, and distribute via the web, a 3D digital library of 1,000 face type examples to test algorythms, and develop initial standards for true 3D facial comparison. The result will be algorythms to quickly and accurately recognize distinctive facial features, and to viably recognize and authenticate indivuals against the database.
