Two Novel Face Recognition Approaches

Abstract

Face recognition represents a very important biometric domain, the human face being a psychological biometric identifier that is widely used in person authentication. Also, it constitutes a popular computer vision domain, facial recognition being the most successful application of object recognition. Recognizing of faces is a task performed easily by humans but it remains a difficult problem in the computer vision area. Automated face recognition constitutes a relatively new concept, having a history of some 20 years of research. Major initiatives and achievements in the past decades have propelled facial recognition technology into the spotlight (Zhao et. al, 2003). A facial recognition system represents a computer-driven application for automatically authenticating a person from a digital image, using the characteristics of its face. As any biometric recognition system, it performs two essential processes: identification and verification. Facial identification consists in assigning an input face image to a known person, while face verification consists in accepting or rejecting the previously detected identity. Also, facial identification is composed of a feature extraction stage and a classification step. Face recognition technologies have a variety of application areas, such as: access control systems for various services, surveillance systems and law enforcement (Zhao et. al, 2003).. Also, these technologies can be incorporated into more complex biometric systems, to obtain a better human recognition. Facial recognition techniques are divided into two major categories: geometric and photometric methods. Geometric approaches represent featurebased techniques and look at distinguishing individual features, such as eyes, nose, mouth and head outline, and developing a face model based on position and size of these characteristics. Photometric techniques are view-based recognition methods. They distill an image into values and compare these values with templates. Many face recognition algorithms have been developed in the last decades. The most popular techniques include Eigenfaces (Turk & Pentland, 1991, Barbu, 2007), Fisherfaces (Yin et. al, 2005), Linear Discriminant Analysis (LDA), Elastic Bunch Graph Matching (EBGM), Hidden Markov Models (HMM) (Samaria et. al, 1994) and the neuronal model Dynamic Link Matching (DLM) (Wiskott et. al, 1996). In this chapter we present two facial recognition approaches. The first one is an Eigenfacebased recognition technique, based on the influential work of Turk and Pentland (Turk & Pentland, 1991). Proposed in 1991 by M. Turk and A. Pentland, the Eigenface approach was the first genuinely successful system for automatic recognition of human faces, representing