Human Identification Based on Segmentation and Frequency Analysis of Cardiac Sounds

1 The performance of traditional biometric identification systems is as yet unsatisfactory in certain 2 fields of application. For this reason other physiological or behavioral characteristics are recently be3 ing considered, using new electrical or physical signals linked to a person’s vital signs. This paper 4 examines the biometric characteristics of PCG (PhonoCardioGram) signals from cardiac auscultation. 5 More specifically, the paper proposes a preliminary study related to the identification of individuals via 6 frequency analysis of cardiac sounds. The results, obtained using a database containing several heart 7 sound recordings from 20 different people, confirm the biometric properties of PCG signals, which can 8 thus be included among the physiological signs used by an automatic identification system. 9 Francesco Beritelli and Salvatore Serrano, Dipartimento di Ingegneria Informatica e delle Telecomunicazioni Università degli Studi di Catania viale Andrea Doria, 6 95125 Catania Italy March 30, 2007 DRAFT