Abstract:
Gender classification is a basic function of determining gender by facial features
which determine gender based on face images. Visual information from faces provides one
of the more important sources of information for gender classification. This system
proposes a gender classification system by faces from Myanmar National Registration
Cards. In this system, Principal Component Analysis (PCA) and Support Vector Machines
(SVM) are used to classify gender from the facial image. PCA is a dimensionality reduction method that is used to reduce the dimensionality of large data sets, by
transforming a large set of variables into a smaller one. SVM is a binary classification that
incorporates PCA in the form of features, which can be predicted from two possible classes.
The face regions were initially detected using the viola jones method, and then the faces
were extracted. Then PCA is performed on the face region for feature extraction to encode
the second-order statistics of the data. These principal components are fed as input to the
SVM for classification. The proposed method is implemented by using the collected dataset
file. The classification rate of the proposed system is described by three datasets; they are
only female images, only male images, and combined male and female datasets for gender
classification. It achieved better performance on all three types of experimentation in this
system. The performance of the system is tested using the own data file with 180 images
(80 males and 100 females) captured from the side of frontal views. SVM classifier
achieves as high as 92.4% gender classification accuracy for 180 input images.
