Securing handheld devices and fingerprint readers with ECG biometrics

This paper investigates the feasibility of collecting viable biometric electrocardiogram (ECG) signals from fingertips. A system for biometric recognition on handheld electronic devices is proposed and analyzed from a flexibility and a permanence point of view. The recognition algorithm is based on the previously proposed Autocorrelation/ Linear Discriminant Analysis (AC/LDA) [3]. In order to assess the flexibility of the acquisition procedure from the fingertips, various electrode configurations are examined along with signal permanence over different signal collection periods. The experimental results indicate very promising error rates (EER of 8.7% over 22 subjects), for ECG data obtained when touching a sensing surface with the fingertips. The envisioned applications include security on handheld electronic devices, such as smartphones and smart cards, as well as multimodal systems which combine ECG and fingerprint biometric readers.