Biometric Identification System Based on Pupillary Hippus: a Preliminary Study

Abstract

Automatic user verification/identification algorithms are an imperative necessity within the field of security systems. The design of more reliable and robust options represents a broad research aim. Biometric systems emerge as a solution for such requirement, but these are not immune to attacks. In the present paper, a biometric solution based on analysis of spontaneous pupillary oscillation (hoppus) are submitted, highlighting its robustness against impostors and spoofs, due to the property of liveness. Throughout the text we describe the methodology for obtaining the information (about a pilot group of seven persons), as well as the proposed feature set for classification purpose. Results for selected classifiers reach maximum values of 100% of training and 85.7% in testing.