AI-Driven Scalable Authentication Framework Using ECG and EEG Biometrics for Enhanced Digital Security

Advancing Internet of Things (IoT) and metaverse innovations require reliable and expandable user authentication systems. Addressing this, the study introduces a Siamese neural network model incorporating Electrocardiogram (ECG) and Electroencephalogram (EEG) biometrics for user authentication. This research stands on the hypothesis that diverse, large-scale datasets are more effective for user authentication than extensive data per individual, emphasizing the importance of generalization over memorization. Utilizing datasets like ECG-ID and PTB, which vary in user count and sample size, the model demonstrates the significance of balancing user diversity with sample number. The findings reveal a model's enhanced ability to generalize to new users without significant accuracy loss, marking a change from common models that tend to overfit with increased familiarity to trained data. This study highlights the potential of EEG and ECG biometrics in developing scalable, accurate authentication systems adaptable to new users, thus enhancing security in digital environments.