Facial Expression Prediction using Information Theoretic Learning

Understanding human emotions is a key area of research. The area of facial expressions is become popular because certain facial expressions have universal meaning, and these emotions have been documented for tens and even hundreds of years. The aim of this work, is to identify the expression reflected by the face of a single person. This is a classification problem that consists of six different expressions of a person. In this work, we have used FERA-2013 dataset which have six different type of expression of individual person. The Linear Binary Pattern (LBP) is used for feature extraction from the face images. Due to the high dimensionality of the data, we have used PCA feature extraction approach for dimensionality reduction. After that, SVM, LMS, KLMS, MCC and KMCC classification approach is used to compute the classification accuracy of the dataset.