ECG-Based Biometric Recognition: A Survey of Methods and Databases.

Area of Science:

• Biometrics
• Signal Processing
• Computer Science

Background:

• Electrocardiogram (ECG) signals offer unique physiological characteristics for biometric identification.
• Existing literature on ECG-based biometrics lacks a consolidated overview of studies and data sources.
• Advancements in signal processing and machine learning enable sophisticated biometric recognition using ECG data.

Purpose of the Study:

• To provide a chronologically ordered survey of studies on ECG-based biometric recognition systems.
• To systematically review and compile the main ECG features and recognition techniques used in the literature.
• To identify, characterize, and compile available ECG databases for biometric research.

Main Methods:

• Conducted a comprehensive literature review of ECG biometric recognition studies.
• Categorized and analyzed ECG features and recognition methodologies.
• Surveyed and documented characteristics of publicly available ECG databases.

Main Results:

• Identified and detailed a wide range of ECG features and recognition techniques.
• Compiled a comprehensive list of relevant research papers and their methodologies.
• Provided a detailed overview of ECG databases, including their key characteristics and usage in studies.

Conclusions:

• This work presents the most complete overview to date of studies and data sources for ECG-based biometrics.
• Researchers can use this survey to understand the state-of-the-art and identify relevant literature and databases.
• Facilitates the testing and validation of novel biometric algorithms using established ECG datasets.