Emotion Classification Based on Single Electrode Brain Data: Applications for Assistive Technology

• Type: Article in International Conference Proceedings Book
• Date: 2023

Authors

Rodrigues D. (Author) Other; reis, lp (Author) FEUP; Faria B.M. (Author) Other

Conference proceedings International

Title: Classification and Data Science in the Digital Age

Pages: 323-331

Indexing

Scopus - 0 Citations

Other information

Authenticus ID: P-00Z-HXJ

DOI: 10.1007/978-3-031-09034-9_35

Abstract (EN): This research case focused on the development of an emotion classification system aimed to be integrated in projects committed to improve assistive technologies. An experimental protocol was designed to acquire an electroencephalogram (EEG) signal that translated a certain emotional state. To trigger this stimulus, a set of clips were retrieved from an extensive database of pre-labeled videos. Then, the signals were properly processed, in order to extract valuable features and patterns to train the machine and deep learning models. There were suggested 3 hypotheses for classification: recognition of 6 core emotions; distinguishing between 2 different emotions and recognising if the individual was being directly stimulated or merely processing the emotion. Results showed that the first classification task was a challenging one, because of sample size limitation. Nevertheless, good results were achieved in the second and third case scenarios (70% and 97% accuracy scores, respectively) through the application of a recurrent neural network.

Language: English

Type (Professor's evaluation): Scientific

No. of pages: 8