Biometric security based on ECG

Recently the electrocardiogram (ECG) has been proposed as a novel biometric. This paper aims to construct a reliable ECG verification system, in terms of privacy protection. To this end, an improved expression to estimate the capacity in the autocorrelation (AC) of the ECG is derived, which not only gives an upper bound for the number of identifiable subjects using ECG, but also helps to determine the number of independent features needed for verification. This capacity derivation, to our best knowledge, is new and can be easily extended to other biometrics. In addition, we apply the Quantization Index Modulation (QIM) scheme to the verification system for template protection. Based on our database, we found that, the capacity in the AC of ECG is approximately 20 bits information contained in the 23 independent features. We have also shown that compared to unprotected systems, the QIM scheme greatly enhances privacy protection of ECG templates without significant sacrificing of the identification performance.