Facial Expression Recognition Using Patch-Based LBPS in an Unconstrained Environment

Facial expression recognition in the wild is challenging due to various unconstrained conditions. Although existing facial expression classifiers have been almost perfect on analyzing constrained frontal faces, they fail to perform well on partially occluded faces common in the wild. In this paper, an improved facial expression recognition technique, patch-based multiple local binary pattern (LBP) descriptor, comprises three and four patch LBPs [TPLBP, FPLBP]. The two-dimensional discrete cosine transform (DCT) was applied over the entire coded TPLBP and FPLBP face image as a feature extractor. The experiment results show that the proposed technique achieves a better recognition rate than state-of-the-art techniques. Oulu-CASIA dataset facial expression images have been evaluated using a support vector machine (SVM) classifier resulted in an accuracy of 92.1%.