Robust ECG based person identification system

Sidek, K 2013, Robust ECG based person identification system, Doctor of Philosophy (PhD), Computer Science and Information Technology, RMIT University.

Document type: Thesis
Collection: Theses

Attached Files
Name Description MIMEType Size
Sidek.pdf Thesis application/pdf 5.04MB
Title Robust ECG based person identification system
Author(s) Sidek, K
Year 2013
Abstract Identity theft is a burgeoning issue. Gaining unauthorized access to computer network tends to compromise the system which could potentially cause undetected fatal destruction and disastrous consequences for individuals and the nation. It is to the extent of taking down communication networks, paralyzing transportation systems and crippling power grids. If security system are burdensome, people may avoid using them, preferring functionality and convenience. For these reasons, an effective security mechanism needs to be deployed in combating identity crimes. Therefore, this thesis proposes of implementing biometric technology as a viable solution for the aforementioned problems.

In the recent years, the electrocardiogram (ECG) signal was introduced as a potential biometric modality to overcome issues of currently available biometric attributes which could be falsified by gummy fingerprints, static iris and face images, voice mimics and fake signatures. When a person is having a heartbeat, automatically it proclaims that the person exist and is alive. Thus, the advantage as a life indicator mechanism verifies the presence of a person during the time of recognition. For the past decade, preliminary investigations on the validity of using ECG based biometric have been manifested with different person recognition methods to support its usability in security and privacy applications. Even though, ECG based biometric has set its ground in recognizing people, however, the underlying issues that governs a practical biometric system have not been properly addressed. Basic problems which require further attention are fundamental issues which touch the aspects of reliability and robustness of an ECG based biometric system in a real life scenario. Thus, in this thesis, we have identified four main research problems which are essentially important to increase user acceptability of ECG based biometric recognition covering different aspects of a practical biometric system such as distinctiveness, permanence, collectability and performance. The research issues being posed in this thesis are the selection of extracted biometric features, subject recognition with different pathological and physiological conditions, performing biometric with low sampling frequency signals and applying ECG based biometric in mobile surroundings.

This thesis suggests of solving ECG based biometric recognition raised problems in a holistic perspective which does not limit the implementations to certain groups of users but looking at the issue as a whole and in a boarder avenue so that it could be applicable to almost all walks of life. A single optimum biometric system that supersedes others does not exist as each biometric modality is based on the nature of the implementation and application. Nevertheless, ECG based biometric features give a strong indication that it would be well accepted by users in the future due to the automatic liveness detection factor which is available in every human being that further expands to people with disabilities such as amputees and those who are visually impaired. Therefore, this thesis is substantial and vital as to assist and provide alternative person identification mechanism to present security and privacy applications in the quest to combat identity crimes.
Degree Doctor of Philosophy (PhD)
Institution RMIT University
School, Department or Centre Computer Science and Information Technology
Keyword(s) ECG biometric
QRS complex
physiological variability and pathological conditions
Version Filter Type
Access Statistics: 262 Abstract Views, 661 File Downloads  -  Detailed Statistics
Created: Fri, 10 Oct 2014, 12:11:54 EST by Maria Lombardo
© 2014 RMIT Research Repository • Powered by Fez SoftwareContact us