Auth ‘n’ Scan

Abstract

Recent commodity smartphones have biometric sensing capabilities, allowing their daily use for authentication and identification. This increasing use of biometric systems motivates us to design an opportunistic way to sense user's additional physiological or behavioral data. We define this concurrent physiological or behavioral data sensing during biometric authentication or identification as dual-purpose biometrics. As an instance of dual-purpose biometrics, we develop photoplethysmography (PPG) sensing during mobile fingerprint authentication, called Auth ‘n’ Scan. Our system opportunistically extracts cardiovascular information, such as a heart rate and its variability, while users perform phone unlock of a smartphone. To achieve this sensing, our Auth ‘n’ Scan system attaches four PPG units around a fingerprint sensor. The system also performs noise removal and signal selection to accurately estimate cardiovascular information. This paper presents the hardware implementation and signal processing algorithm of our Auth ‘n’ Scan prototype. We also report our system evaluations with 10 participants, showing that, despite a little low precision (a standard deviation of 3--7), estimation of heart rates with high accuracy (under a mean error of 1) is possible from PPG data of five seconds and longer if their baseline information is given. We discuss the feasibility of opportunistic PPG sensing in mobile fingerprint authentication.