Local Binary Patterns Based Facial Expression Recognition for Efficient Smart Applications

Abstract

Facial expressions are direct means of communication of human’s emotional state. Hence facial expression recognition (FER) has always been a topic of great interest of researchers specially for smart applications. Numerous approaches have been proposed for FER using Local binary patterns (LBP). This chapter presents an analysis of LBP feature descriptor for FER. State of the art approaches have been discussed that use LBP as their main component. Here, basic LBP operator along with several variants and their main properties are described that have been proved useful for FER. A general framework for FER is described which includes four consecutive modules. These modules are preprocessing, feature extraction, dimensionality reduction and classification. LBP based FER results have been reported on three benchmark datasets JAFFE, CK+ and Yale. Experimentation demonstrates usefulness of LBP in FER.