A multiband approach to human identity verification based on PhonoCardioGram signal analysis

Abstract

New physiological or behavioural characteristics have to be considered in order to improve the performance of automatic mono/multi-biometric recognition systems. It has recently been shown in [1], by means of a preliminary study, that PhonoCardioGram (PCG) signals have specific individual characteristics that can be taken into consideration as a physiological sign used in a biometric system. This paper presents new results in biometric identification/verification via frequency analysis of cardiac sounds. More specifically, the paper proposes a multiband approach to PCG signals that has made it possible to enhance the linear separability and obtain an Equal Error Rate (EER) of 9 %.

PCG signals for biometric authentication systems: An in-depth review

“…They worked on 10 user's achieving an accuracy of 96%. Beritelli et al, [21] developed an approach based on sub-band aggregation for verification. The PCG signal is segmented to obtain S1 and S2 from the PCG signal.…”

“…Other approaches were based on dividing the PCG signal into frames. Then an autocorrelation function was applied for ordering and determining the variant periodic participles of the energy signal [19,21,22,28,29,39]. Multiplication of the frame values by a hamming window was performed for minimizing the discontinuation disruption at the start and the end of each frame.…”

“… The main aim of this framing is to create independence between samples and to perform an accurate estimate.  It helps in finding the variance of each segment and to increase the effect of the segment parts S1 and S2 and enhance them[19,21,22,28,29,39],…”

Heart Sound Biometric System Based on Marginal Spectrum Analysis

“…The system's performance in terms of accuracy identification rate could reach more than 95%, but the experimental sample was inadequate because the total sample size was only 10 [7]. Another method was based on an identification algorithm of heart sound's Fourier spectrum, and the results show that the performance in terms of equal error rate (EER) can be reduced to 9% [8,9]. …”