Recurrent Neural Networks (RNNs) to improve EEG-based person identification
Youssef Mohamed, Ahmed M. Anter, Ahmed B. Zaky
Abstract
This study explored the potential of using EEG signals for biometric identification, and the performance of advanced deep learning models, such as LSTM and GRU, in extracting features from raw EEG signals was evaluated. It was demonstrated that these models could accurately identify individuals based on their unique brain activity patterns. The study’s findings have significant implications for secure and reliable biometric identification systems in healthcare and security. However, challenges remain in addressing data privacy and security, cross-domain generalization, and fairness and bias. Further research is needed to address these challenges and fully realize the potential impact of deep learning on biometric identification.