An RFID-Based BCI System for Emotion Detection Using EEG patterns

Emotion categorization using electroencephalogram (EEG) signals can be a difficult task and requires cognitive analysis of EEG patterns for mapping them to a specific emotion. Such a system can be utilized in many ways. The concept of emotions is itself complex and fuzzy in nature. Therefore, to handle the fuzziness of individual’s emotional experiences and presenting them as a discrete emotion is all that is required by many EEG-based brain computer interface (BCI) systems. In this work, we present an RFID-based BCI system for Emotion classification from EEG patterns. The presented system uses a fuzzy inference system (FIS) that utilizes the valence and arousal modalities to model the individual’s emotion. The major steps of the system are: preprocessing recorded EEG patterns and saving them with individual’s identification using RFID tags, then preprocessed EEG patterns are decomposed into different frequency bands using wavelet decomposition, followed by feature extraction step, feature matrix so obtained is used for training a k-nearest neighbor model to generate valence and arousal values for given EEG patterns. The output of KNN is fed to the FIS which models the pattern as a discrete emotion. system gives a new dimension to the BCI-based content servers and can also be used in healthcare services which is the need of the hour.