EEG Channel Correlation Based Model for Emotion Recognition

Emotion recognition using Artificial Intelligence (AI) is a fundamental prerequisite to improve Human-Computer Interaction (HCI). Recognizing emotion from Electroencephalogram (EEG) has been globally accepted in many applications such as intelligent thinking, decision-making, social communication, feeling detection, affective computing, etc. Nevertheless, due having too low amplitude variation related time on EEG signal, the proper of this signal become challenging. Usually, considerable effort required identify feature or set for an effective feature-based system. To extenuate manual human extraction, we proposed deep machine-learning-based model with Convolutional Neural Network (CNN). At first, one-dimensional data were converted Pearson's Correlation Coefficient (PCC) featured images channel correlation sub-bands. Then fed into CNN recognize emotion. Two protocols conducted, namely, protocol-1 two levels and protocol-2 three valence arousal that demonstrate We investigated only upper triangular portion PCC reduced computational complexity size memory without hampering accuracy. The maximum accuracy 78.22% 74.92% obtained internationally authorized DEAP dataset.