| Abstract |
Real-time EEG-based emotion recognition may now be accomplished with commercially available, low-cost, portable EEG equipment. One characteristic that these devices have is a limited number of channel options. Consequently, the subsequent procedure will be affected by the decision of where to connect the electrode in the circuit. Online schooling is one of the most recent real-time uses of EEG-based emotion identification technology. The feelings of learners are extremely important to e-learning systems since such feelings may either stimulate or inhibit learning. Through the utilization of a deep neural network, it is possible to extract a sizeable quantity of information from a constrained number of EEG channels. Most of the time, a particular area of the brain is in charge of a particular function; nevertheless, these duties are subject to change depending on the problem that must be handled. In the investigation, we make use of two strategies that have shown themselves to be highly effective in order to generate a feature selection that develops throughout the study. The system makes use of primarily distributed and centralized approaches for the selection of features in equal measure. Because the requirements for an application can change depending on the problem that is being attempted to be solved, those problems place limitations on the response time, resources, and accuracy level. A variety of feature selection methods that feed into a variety of classifiers can be used to strike a balance between the various application restrictions. © 2022 IET Conference Proceedings. All rights reserved. |
