Review of fiducial and non-fiducial techniques of feature extraction in ECG based biometric systems

Abstract

Background/Objectives: Biometric is a tool of measuring and statistically analyzing biological data. Due to advancement in technology, spoofing attacks and credential forgery are becoming very common issues of modern societies. Methods/Statistical Analysis: Over the last decade, the Electrocardiogram (ECG) is known as an emerging biometric instrument for individual identification and verification as the ECG varies among people because of their diverse anatomy of the heart. At present ECG is a popular research topic in the area of physiological biometrics. The greater part of ECG biometric literature employs fiducially based features, resulting from spikes, crest and temporal marker of ECG signal. Findings: The main focus of this review is to provide scientific analysis and comparison between fiducial and non-fiducial techniques of feature extraction especially in terms of efficiency in large and small datasets. It also provides a key manifestation of future research perspectives in the field of ECG based biometrics. Application/Improvements: The proposed review can be useful in further research in the same area.