Facial expression recognition using PCA and texture-based LDN descriptor

Abstract

The most expressive way in which humans can display their emotion is through facial expression. The development of automated system that performs this task is very difficult. In this paper, we introduced texture- based LDN descriptor along with PCA which helps to recognize facial expression in efficient manner. In order to obtain directional information, we are using two masks i.e. Kirsch and Robinson mask and we compare their efficiency. Experiments shows that this method is more efficient compared to existing works.