Electrocardiogram Based Identification using a New Effective Intelligent Selection of Fused Features.

Area of Science:

• Biometrics
• Signal Processing
• Artificial Intelligence

Background:

• Human identification using Electrocardiogram (ECG) signals is an active research area.
• Existing methods for ECG-based identification face challenges in feature selection and accuracy, especially with arrhythmias.

Purpose of the Study:

• To propose a novel and effective intelligent feature selection method for human identification using ECG signals.
• To enhance the accuracy and robustness of ECG-based identification systems.

Main Methods:

• ECG signal preprocessing and feature extraction.
• Feature compression using cosine transform.
• Intelligent feature selection via a hybrid approach combining genetic algorithms and artificial neural networks.

Main Results:

• The proposed method achieved high identification rates across three public ECG databases: 99.89% (MIT-BIH Arrhythmias), 99.84% (MIT-BIH Normal Sinus Rhythm), and 99.99% (European ST-T).
• Remarkable identification accuracy was demonstrated for both normal ECG signals and those with various arrhythmias.
• High performance was achieved with a reduced set of selected features.

Conclusions:

• The proposed intelligent feature selection method is highly effective for human identification using ECG signals.
• The algorithm demonstrates robustness and high accuracy in identifying subjects from ECG data, irrespective of arrhythmias.
• This method offers a promising solution for secure and accurate biometric identification systems based on ECG.