Occlusion invariant face recognition system

Abstract

Face recognition has acquired a lot of attention in market and research communities, but still remained very accosting in real time applications. It is one of the several techniques used for identifying an individual. In face recognition system there are many factors which affect the performance of a system. The major factors affecting the face recognition system are pose, illumination, ageing, occlusion and expression etc. Among the above mentioned problem an occlusion is most affecting problem in face recognition. In a face recognition system due to obstacles like sunglasses, scarf etc. we cannot recognize a face image. So first we detect an occlusion from a face image by using a SVM (Support Vector Machine) classifier. To resolve the occlusion problem, each face is divided into k local regions which are analyzed in isolation. We discard an occluded part in a face image and based on remaining non-occluded part of a face image we will recognize a face image. For face recognition purpose we will be using a near set theory.