Paper Key : IRJ************647
Author: Maaz Alam
Date Published: 01 Jan 2025
Abstract
EEG-based biometric identification has emerged as a highly reliable and non-invasive modality for individual recognition through the evaluation of unique patterns of brain activity. While promising, getting solid performance and high accuracy is still very difficult. So, this research aims at the feature extraction improvement and utilizes Siamese convolutional 1D neural networks to detect individual differences based on frequency patterns in EEG data. After rigorous investigation, two experimental configurations were identified: The model was trained on a single task and tested on multiple tasks with the ability to generalize across different brain activities. The following analysis showed that the recognition accuracy was above 99%, which emerged the effectiveness of the proposed method. This study improves upon the existing work on EEG biometric systems by presenting one that is both highly efficient and secure; hence, it demonstrates the power of deep learning for creating novel identification systems.
DOI Requested