A Novel Method for Gender and Age Detection
Based on EEG Brain Signals
Haitham
Issa1, Sali Issa2, and Wahab Shah3
1Department of Electrical Engineering, Zarqa
University, Jordan
2Department of Computer Engineering, Hubei
University of Education, China
3Department of Electrical Engineering, Namal
University, Pakistan
Abstract: This
paper presents a new gender and age classification system based on Electroencephalography (EEG) brain
signals. First, Continuous Wavelet Transform (CWT) technique is used to get the
time-frequency information of only one EEG electrode for eight distinct
emotional states instead of the ordinary neutral or relax states. Then,
sequential steps are implemented to extract the improved grayscale image
feature. For system evaluation, a three-fold-cross validation strategy is
applied to construct four different classifiers. The experimental test shows
that the proposed extracted feature with Convolutional Neural Network (CNN)
classifier improves the performance of both gender and age classification, and
achieves an average accuracy of 96.3% and 89% for gender and age
classification, respectively. Moreover, the ability to predict human gender and
age during the mood of different emotional states is practically approved.
Keywords: EEG, Gender, Age, CWT, CNN.
Received September 27, 2020; accepted February 9, 2021