FACIAL EXPRESSION ANALYSIS USING DIFFERENCE THEORETIC TEXTURE FEATURES

Abstract

Human beings are able to detect facial expressions without any effort or trouble while it is a tedious task for the computer. Therefore, only developing an automated system that is powerful and able to analyse the facial expressions is not the lone job. There are several other issues that need to be addressed: extracting face from an image, facial feature extraction which provides information about the facial expressions of the person. These systems need the capacity to perceive and understand emotions as communicated by the facial muscles. Seeing this automated analysis of non-verbal behaviour, many researchers have become interested in this field. An automated system that is able to perform all the above actions forms a major leap in achieving computer-like interaction. This thesis deals and gives a solution to the above problem. An attempt has been made to provide an optimal and simple solution to achieve Facial expression recognition and an analysis has been made taking different algorithms and techniques and classifiers into account to get the best result. Not only, do we discuss about a good facial feature which would take account of every detail in the face but at the same a good classifier is needed to correctly classify the facial expression.