Optimized Artifical Neural Network For EEG-Based Emotion Recognition

The Emotion provides a great impact in cognitive activities of the humans, such as communication and interaction. The emotion recognition is the hot research topic as the emotion plays a significant role in artificial human computer interaction. The EEG is considered as the best modality considered for the emotion recognition, as it directly receives the signal from the human brain. Hence, in this research a novel optimization for the efficient emotion recognition based on the EEG signals is proposed. The classification of the emotion process is carried out by the Optimized Artificial Neural Network, in which the Brain storm Optimization (BSO) is employed for tuning the hyper parameter. Further, the process, such as artifacts removal, feature extraction, electrode selection enhances the accuracy of the classifier. The experimental results show that the proposed EEG based emotion recognition process attains maximum accuracy, sensitivity and specificity of 94.23% , 93.83% and 95.60% respectively.