Increasing the accuracy of ECG based biometric analysis by data modelling

Abstract Here an effort is made to use human electrocardiogram as a tool of biometric analysis for authentication. The proposed method is based on first accurate extraction of characteristic features from each ECG and then design of a suitable classification methodology to comment on the authenticity. As the feature matrix is a huge one, Principal Component Analysis (PCA) is applied to avoid handling of large amount of data. Next, the reduced features from PCA are fitted into a quadratic polynomial model by the method of least square. Then the fitted values for the allowed set of data is obtained and the range over which they vary, provides the signature matrix of a person. Finally the classification is done by a comparison based on nearest neighbor method. The method is tested on ECG of 20 individuals taken from PTB database. This method has accuracy more than 95% with the best fit modeling which becomes only 80% without data modeling proving the importance of best fit modeling of data before classification. This accuracy is comparable with conventional biometric techniques; moreover, ECG biometric can be used with other authentication scheme, with ECG providing liveliness proof.

[1]  P. Gupta,et al.  ECG to Individual Identification , 2008, 2008 IEEE Second International Conference on Biometrics: Theory, Applications and Systems.

[2]  Adrian D. C. Chan,et al.  Wavelet Distance Measure for Person Identification Using Electrocardiograms , 2008, IEEE Transactions on Instrumentation and Measurement.

[3]  M. Kyoso.  A technique for avoiding false acceptance in ECG identification , 2003, IEEE EMBS Asian-Pacific Conference on Biomedical Engineering, 2003..

[4]  W. J. Tompkins,et al.  Estimation of QRS Complex Power Spectra for Design of a QRS Filter , 1984, IEEE Transactions on Biomedical Engineering.

[5]  Adriaan van Oosterom,et al.  Geometrical aspects of the interindividual variability of multilead ECG recordings , 2001, IEEE Transactions on Biomedical Engineering.

[6]  A. Uchiyama,et al.  Development of an ECG identification system , 2001, 2001 Conference Proceedings of the 23rd Annual International Conference of the IEEE Engineering in Medicine and Biology Society.

[7]  Ola Pettersson,et al.  ECG analysis: a new approach in human identification , 2001, IEEE Trans. Instrum. Meas..

[8]  Ognian Boumbarov,et al.  ECG personal identification in subspaces using radial basis neural networks , 2009, 2009 IEEE International Workshop on Intelligent Data Acquisition and Advanced Computing Systems: Technology and Applications.

[9]  F. Sufi,et al.  Legendre Polynomials based biometric authentication using QRS complex of ECG , 2008, 2008 International Conference on Intelligent Sensors, Sensor Networks and Information Processing.

[10]  Chih-Yu Hsu,et al.  A Novel Personal Identity Verification Approach Using a Discrete Wavelet Transform of the ECG Signal , 2008, 2008 International Conference on Multimedia and Ubiquitous Engineering (mue 2008).

[11]  Brenda K. Wiederhold,et al.  ECG to identify individuals , 2005, Pattern Recognit..

[12]  Paul Kligfield,et al.  The centennial of the Einthoven electrocardiogram. , 2002, Journal of electrocardiology.

[13]  Dimitrios Hatzinakos,et al.  Analysis of Human Electrocardiogram for Biometric Recognition , 2008, EURASIP J. Adv. Signal Process..

[14]  Hsiao-Lung Chan,et al.  Human identification by quantifying similarity and dissimilarity in electrocardiogram phase space , 2009, Pattern Recognit..

[15]  Rosli Besar,et al.  A New Approach to ECG Biometric Systems: A Comparitive Study between LPC and WPD Systems , 2010 .

[16]  I. Jolliffe.  Principal Component Analysis , 2002 .

[17]  Yu Hen Hu,et al.  One-lead ECG for identity verification , 2002, Proceedings of the Second Joint 24th Annual Conference and the Annual Fall Meeting of the Biomedical Engineering Society] [Engineering in Medicine and Biology.

[18]  John J. Soraghan,et al.  ECG based biometric for doubly secure authentication , 2011, 2011 19th European Signal Processing Conference.