Feature regularization and extraction in eigenspace for face recognition

Abstract

The ability to recognize human faces is a demonstration of incredible human intelligence. Over the last two decades researchers from diverse areas are making attempts to replicate this outstanding visual perception of human beings in machine recognition of faces. Within the face recognition literature, debate has been centered on how human beings perceive human faces and this has become an important and active research area. Psychologists concluded that holistic and component based approaches are dual routes to the face recognition. Recent studies (like FERET competition) show that holistic approaches have dominated the face recognition systems and have shown better performance than omponent based approaches. Although these holistic/appearance based approaches have attained certain level of maturity their performances are far away from the abilities of human recognition of faces. Owing to the immense potentiality of the face recognition applications it is imperative to develop a face recognition system, which is robust, e±cient and able to achieve high recognition accuracy on large face image databases. In this thesis, we propose various algorithms which are based on statistical pattern recognition and computer vision for robust face recognition with high accuracy.