High Confidence Multimodal Biometric System

Rapid and accurate identification of all individuals has been the goal of the biometric industry for many years. Unfortunately, due to limitations in technology and variations in the biometric itself, no one biometric imaging modality has been able to satisfy the performance requirements of most large scale identity management applications. This research addresses the fundamental hypothesis of whether two or more biometric systems can be combined to deliver improved performance over a single biometric system. This work will develop the scientific testing methodology to combine multiple biometric measurements in order to achieve superior performance over a single biometric system. A general purpose biometric testing tool will be developed that will accept enrollment and inquiry data from multiple biometric imaging systems and generate the corresponding ROC performance curves. Upon successful completion of this initiative, system performance requirements can be presented in the form of Pfm, Pfnm, and throughput, and a metric driven approach taken to determine if those requirements are in fact, achievable. The system designer can then ascertain what types of biometric technology would be required to achieve the desired performance, and how the individual biometric technologies should be combined to meet the stated goal.