Hardware and software development of Ecg biometric system
Loading...
Date
2018-06
Authors
Chia, Pei Kiak
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Security is a top issue in the world and nothing is completely secure. Now biometrics
has been a useful security tool in recent decades because of its effectiveness in identifying
and authenticating desired users versus potential intruder. ECG has been proposed as
promising biometrics for human identification and authentication. The advantage of ECG
biometric is highly secure, and confidential compared to fingerprint, iris, and face.
However, 12 number of leads used to collect ECG signal in typical identification system
is not efficient. So that, this project proposes a portable ECG detection kit that is designed
by using solely by Lead 1 for ECG detection and recording. The Android application also
helps to collect proper ECG signal by checking the status of subject’s ECG in this project
before the subject’s ECG data is recorded. In this project, Pan-Tompkins algorithm is
used as feature extraction technique of ECG fiducial points while artificial neural network
is used for identification system. The performance evaluation shows that 93.65% of
overall system performance and 85% of individual subject accuracy based on 10 different
subject’s training data and 20 different subject’s testing. This experimental result is
promising hence ECG signal is proved to be a viable modality for biometric system.