Impact of the acquisition time on ECG compression-based biometric identification systems

Abstract

The ECG signal conveys desirable characteristics for biometric identification (universality, uniqueness, measurability, acceptability and circumvention avoidance). However, based on the current literature review, there are no results that evaluate the number of heartbeats needed for personal identification. This information is undoubtedly useful when building a biometric identification system – any system should ask participants to provide data for identification, using the smallest time interval that is possible, for practical reasons. In this paper, we aim at exploring this topic using a measure of similarity based on the Kolmogorov Complexity, called the Normalized Relative Compression (NRC). To attain the goal, we built finite-context models to represent each individual – a compression-based approach that has been shown successful for several other pattern recognition applications like image similarity, DNA sequences or authorship attribution.