Face recognition techniques, their advantages, disadvantages and performance evaluation

Abstract

A human brain can store and remember thousands of faces in a person's life time, however it is very difficult for an automated system to reproduce the same results. Faces are complex and multidimensional which makes extraction of facial features to be very challenging, yet it is imperative for our face recognition systems to be better than our brain's capabilities. The face like many physiological biometrics that include fingerprint, hand geometry, retina, iris and ear uniquely identifies each individual. In this paper we focus mainly on the face recognition techniques. This review looks at three types of recognition approaches namely holistic, feature based (geometric) and the hybrid approach. We also look at the challenges that are face by the approaches.