Correlation Study of ECG Signals for Biometrics Purposes

Selecting a proper data source for biometrics systems requires it to have very high parameters constancy for a long period of time. Most of the biometrics systems currently in use process fingerprints or iris features which do not change during life of a person. Recent researches propose the use of electric signals (electrocardiogram (ECG), electromyogram etc.) as the main source of data, however these signals are known to change as the human is aging. The developed algorithms that use an ECG show high accuracy in the short term, but studies of the decrease in their effectiveness over time are yet to be conducted. This paper analyzes the short-term (intraday) and medium-term (weeks, months) stability of the electrocardiogram signal based on the correlation analysis of raw QRS complexes and their metrics.