Phonocardiogram Heartbeat Segmentation and Autoregressive Modeling for Person Identification

With the rapid advancement of biosensors and increasing demand of more secured biometric authentication system, cardiac signals are getting special attention. Because of its very simple acquisition technique, phonocardiogram (PCG) signal is getting popularity in this field. This paper presents an au...

Full description

Saved in:
Bibliographic Details
Main Authors: Fahad, Imran, Apu, Md Ashiqur Rahman, Ghosh, Abesh, Fattah, Shaikh Anowarul
Format: Conference Proceeding
Language:English
Subjects:
Autoregresive (AR) Model
Biometric
Burg's Method
Classification
Feature extraction
Heart beat
Hilbert-Huang Transform
Mel frequency cepstral coefficient
Noise reduction
Person Identification
Phonocardiogram (PCG)
Phonocardiography
Transforms
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:With the rapid advancement of biosensors and increasing demand of more secured biometric authentication system, cardiac signals are getting special attention. Because of its very simple acquisition technique, phonocardiogram (PCG) signal is getting popularity in this field. This paper presents an automatic person identification scheme based on autoregressive modeling of PCG beats. On a given PCG recording first preprocesing and then wavelet denoising are applied. In order to perform beat by beat operation, a segmentation scheme is proposed using the Hilbert envelope which extracts the PCG beat containing the first and second heart sounds. Next reflection coefficients are extracted by employing the AR Burg modeling of the PCG beat. Finally the AR Burg reflection coefficients are used in ensemble bag trees classifier to identify a person. Performance of the proposed method is being tested on PCG signals of 50 different person taken from a publicly available PCG dataset and very satisfactory identification performance is achieved.
ISSN:2159-3450
DOI:10.1109/TENCON.2019.8929563