Journal of
Engineering and Computer Innovations

  • Abbreviation: J. Eng. Comput. Innov.
  • Language: English
  • ISSN: 2141-6508
  • DOI: 10.5897/JECI
  • Start Year: 2010
  • Published Articles: 32

Full Length Research Paper

Implementation of a one-lead ECG human identification system on a normal population

Tsu-Wang (David) Shen1*, Willis J. Tompkins2 and Yu Hen Hu3
  1Department of Medical Informatics, Tzu Chi University, Hualien, Taiwan 701, Sec. 3, Jhong-Yang Rd., Hualien, 97004, Taiwan. 2Department of Biomedical Engineering, University of Wisconsin, Madison, WI, USA. 3Department of Electrical and Computer Engineering, University of Wisconsin, Madison, WI, USA
Email: [email protected]

  •  Published: 30 January 2010



The electrocardiogram (ECG) is not only a very useful diagnostic tool for clinical purposes, but also is a potential new biometric tool for human identification. TheECG may be useful as a biometric in the future, since it can easily be combined with other biometrics to provide a liveness check with little additional cost. This research focused on short-term, resting, Lead-I ECG signals recorded from the palms. A total of 168 young college volunteers were investigated for identification as a predetermined group.  Fifty persons were randomly selected from this ECG biometric database as the development dataset. Then, the identification algorithm developed from this group was tested on the entire database. In this research, two algorithms were evaluated for ECG identification during system development. The algorithms included template matching and distance classification methods. Signal averaging was applied to generate ECG databases and templates for reducing the noise recorded with palm ECG signals. When a single algorithm was applied to the development dataset, the identification rate (that is, rank one probability) was up to 98% (49 out of 50 persons). However, when the prescreening process was added to construct a combined system model, the identification rate increased to 100% accuracy on the development dataset. The combined model formed our ECG biometric system model based on results from the development dataset. The identification rate was 95.3% when the same combined system model was tested on the entire ECG biometric database.


Key words: Biometrics, biometric liveness tests, electrocardiogram (ECG), ECG features, identification, template matching, distance classification