A Statistical Approach to Biometric Identity Verification Based on Heart Sounds

One of the most recent innovations in the field of biometric recognition is the usage of heart sounds as physiological traits for identity verification. In this paper, we propose a statistical approach, supported by Gaussian Mixture Models, to the the problem of biometric identity verification based on heart sounds. The system is validated on a database of heart sounds acquired from 147 people, and shows a significant performance improvement with respect to other approaches, yielding an Equal Error Rate of 15.53%.