Provide an EEG Signal-Based Authentication System Using a Hybrid Neural Network
There is potential for biometric uses of EEG signals from the brain. Biometric authentication systems, such as brainwave-based authentication, are extremely significant in our daily lives, and they have several advantages over other approaches. In this study, we investigate the effectiveness of single-channel brainwave authentication systems and identify the most effective channels depending on specific mental activity. Seven persons participated in this study, which employed data from a set of five mental activities (325 samples). Preprocessing, feature extraction, and classification are the three stages of the EEG-based authentication system. It is our hope that the deep layers themselves will execute feature extraction as part of a combined deep learning system that we present in this paper. After passing through multiple deep layers and an LSTM, the raw data output is given to the classifier layer, where the classification process is done. For a single-channel authentication system, we attained an accuracy of 97 to 98 percent by positioning the electrode optimally for mental tasks. In addition, we looked at the authentication system as a whole, and the O2 channel was shown to be 96% accurate in its selection as the best channel. By lowering the number of EEG channels, channel optimization can improve performance and identify the best electrode replacement for different mental processes.