Emotion detection using physiological signals EEG & ECG

• Main Authors: AlzeerAlhouseini, Amjad M.R., Alshaikhli, Imad Fakhri Taha, Abdul Rahman, Abdul Wahab, Dzulkifli, Mariam Adawiah
• Format: Article
• Language: English
• Published: Convergence Information Society(CIS) 2016
• Subjects: QA75 Electronic computers. Computer science
• Online Access: http://irep.iium.edu.my/51279/1/IJACT-amjad.pdf; http://irep.iium.edu.my/51279/; http://www.globalcis.org/dl/citation.html?id=IJACT-3585
• Institution: Universiti Islam Antarabangsa Malaysia

Emotion modeling and identification has attracted substantial interest from disciplines including computer science, cognitive science and psychology. Despite the fact that a lot of qualitative studies have been carried out on emotion, less investigated aspects include the quantifying of physiological signals. This paper presents two physiological signals which are ECG and EEG and shows analysis of its emotional properties. A solution based on the short Fourier transform is proposed for the recognition of dynamically developing emotion patterns on ECG and EEG. Features extraction that are used in this paper are Kernel Density Estimation known as (KDE) and Mel-frequency cepstral coefficients known as MFCC. The classifier that is used in this work is Multi-layer Perceptron known as MLP, classification features are based on the valence and arousal. The experimental setup presented in this work for the elicitation of emotions is based on passive valence /arousal. The results shows that the ECG signal has direct relationship with the arousal factor rather than the valence factor. Also, EEG signal using 19 channels reported high accuracy results for determining emotions.