Performance Evaluation Of Convolutional Neural Network (CNN) For EEG Emotion Classification

Yong, Chun Keong (2018) Performance Evaluation Of Convolutional Neural Network (CNN) For EEG Emotion Classification. Project Report. UTeM, Melaka, Malaysia. (Submitted)

Text (24 Pages) Performance Evaluation Of Convolutional Neural Network (CNN) For EEG Emotion Classification.pdf - Submitted Version Download (474kB)

Abstract

Emotion classification through facial expression or speech intonation is not reliable as human can hide their emotion when expressing their feelings. Therefore, a deep learning technique, Convolutional Neural Network (CNN) is implemented and optimized in this project to analyze human emotion in a more reliable manner. Experimental paradigm is designed by using audio-visual stimuli selected from IAPS and IADS-2 database to acquire EEG data with different emotions. The proposed CNN algorithm is trained on the collected EEG data and then validated by using an open source dataset (SEED). The proposed CNN algorithm achieves the best accuracy of 65% (2 classes of emotion) and 82% (3 classes of emotion) form EEG data collected in the lab and SEED dataset, respectively.

Actions (login required)

View Item