A Fast Fusion Technique for Fingerprint and Iris Spatial Descriptors in Multimodal Biometric Systems

Abstract

Multimodal biometric identification systems aim to combine two or more physical or behavioral traits to provide optimal False Acceptance Rate (FAR) and False Rejection Rate (FRR). In this paper a new fusion technique for fingerprint and iris spatial descriptors with a low execution time is presented. The report focuses on spatial fusion strategies, offering and proposing a modern perspective on multi-biometrics. In greater detail, a spatial-based approach and a homogeneous biometric vector, integrating iris and fingerprint data, are generated, fused, and processed for the overall system matching score. The goal of this approach is to demonstrate that, by using spatial iris features following the standard fingerprint approaches (faster and easier to extract with considerable saving execution time), the recognition system can sustain a comparable robust level with respect to literature frequency based approaches. The resulting multimodal system achieves interesting performance with several commonly used databases, in terms of both accuracy rate and authentication time. As an example, an interesting working point with FAR = 0% and FRR = 7.07%÷9.64% has been obtained, by using the well-known BATH database and the FVC2002 DB2A database. The biometric trait processing time requires 6.44 Sec on a general purpose computer.