Using convolutional neural network to recognize emotions from scalp electroencephalography plot images

Real-time variations in brain activity are determined by electroencephalogram (EEG) data. EEG signals are commonly used in studies to analyze human emotional states. Emotions EEG signals vary from person to person because they each have different emotional responses to the same stimuli. The objective of this study was using EEG signals in emotion recognition.

We specifically focused on employing convolutional neural network (CNN) for detecting image-based emotions in long-term EEG data. After filtering, the EEG data is divided into short sections based on a certain time window and they are converted into EEG plot images. Each of these is classified by convolutional neural networks.

In comparison with the existing methods, the error rate has been reduced and the accuracy rate is better than the existing methods. The mean accuracy of the compared articles is 62.87, 70.50, 74.88, 82.88 and 68.11, but the average accuracy of the proposed method is 85.13.

This research demonstrates the potential and accuracy of CNN in recognizing emotions from scalp EEG plot images. The study contributes to the growing field of emotion recognition and paves the way for future advancements in utilizing CNN for analyzing EEG signals, ultimately aiming to use as an effective method for computer-aided recognition.