The following information was submitted:
Transactions: WSEAS TRANSACTIONS ON BIOLOGY AND BIOMEDICINE
Transactions ID Number: 52-139
Full Name: Mohd Rizon Mohamed Juhari
Position: Associate Professor
Age: ON
Sex: Male
Address: Department of Biomedical Technology, College of Applied Medical Sciences, King Saud University, P. O. Box 10219, Riyadh 11433
Country: SAUDI ARABIA
Tel: 966-1-4693764
Tel prefix:
Fax: 966-1-4693667
E-mail address: mjuhari@ksu.edu.sa
Other E-mails:
Title of the Paper: Asymmetric Ratio based Channel Reduction Method for Assessing the Human Emotions using EEG
Authors as they appear in the Paper: Mohamed Rizon
Email addresses of all the authors: mjuhari@ksu.edu.sa
Number of paper pages: 13
Abstract: Electroencephalogram (EEG) is one of the most reliable physiological signals used for detecting the human emotional changes from the brain. A larger set of 63 EEG channels (or electrodes) in International 10-10 system of arrangement has been adopted in estimating human emotions. However, reduced set of EEG channels in determining human emotion has the advantages of reduced feature size, reduced computational complexity and increased robustness in emotion classification. Selection of reduced channels in emotion classification is the main objective of this paper. In this paper, we propose a method of a new channel selection based on Asymmetric Ratios (Asymmetric Variance Ratio (AVR) and Asymmetric Power Ratio (APR)). This approach is illustrated with 63 EEG channels recorded data for 5 discrete emotions (disgust, happy, surprise, sad, and fear) over 5 healthy subjects. We have considered a set of 28 homogeneous channel pairs from the original set of 63 EEG channels f!
or computing the asymmetric ratios. Through our channel selection algorithm, it is reduced to 4 and 2 channel pairs. The Multi-Resolution Analysis (MRA) based feature extraction is adopted in both original and reduced set of channels for deriving the statistical features (energy, power and normalized energy) from the alpha frequency band. The classification ability of these features was empirically evaluated by using two classifiers such as k-Nearest Neighbor (kNN) and Fuzzy C-Means (FCM) clustering. The efficacy of emotion classification on FCM and kNN is discussed through three performance indices and through classification accuracy, respectively. The maximum mean classification accuracy of 60.557% is achieved using 4 channel pairs among other channel combinations. The maximum individual classification accuracy of; disgust: 68.340%, happy: 63.591%, surprise: 59.893%, sad: 60.458% and fear: 58.892% is achieved though kNN. The paper concludes by discussing the impact of red!
uced channels on emotion recognition.
Keywords: EEG, Asymmetric Ratios, Channel selection, Fuzzy C-Means (FCM) clustering, k-Nearest Neighbor (kNN)
EXTENSION of the file: .pdf
Special (Invited) Session:
Organizer of the Session:
How Did you learn about congress:
IP ADDRESS: 86.51.18.17
