Publication

Abstract : Brain Computer Interface (BCI) enables the direct communication between computers and humans. Event related potentials (ERPs) are created using speller paradigms, such as the P300 wave on user-selected targets. This technique is used in a P300 speller, which allows the user to type characters by detecting P300 waves. The P300 speller is a binary classification problem. The classification is to classify a P300 and Non P300 wave in the EEG signal. Deep learning-based classification of EEG signals using decomposition techniques like Variational Mode Decomposition and Wavelet is used in this work. The proposed approach uses raw signal and decomposed signal for classification of EEG signals recorded from two subjects. The parameters used for evaluation are accuracy, precision, recall, and F1-score. The accuracy of 82% was obtained for the raw signals and the accuracy of 83% was obtained for the decomposed signals using wavelet and variational mode decomposition.