ECG biometrie template protection based on secure sketch scheme

Abstract

ElectroCardioGram ECG biometrics has recently been identified as a promising technique to identify subjects. Meanwhile, as ECG related data can reveal other factors like medical disease, the protection of the ECG biometric template is mandatory. The challenge is to guaranty the privacy of the ECG data, while keeping adequate performance results in terms of false acceptance rate and false rejection rate in subject identification. In this work, the focus will be on designing and assessing a secure sketch scheme for protecting ECG template. We show that the proposed scheme preserves the identification performances of the unprotected system while maintaining an acceptable identification speed compared with the fuzzy commitment scheme. Performance evaluation of the investigated method is reported using data from the physionet database.