A Biometric Security System based on a Hybrid Face Recognition Technique

R. Miragaia (Portugal), M.A. Vega-Rodríguez, J.A. Gómez-Pulido, and J.M. Sánchez-Pérez (Spain)


Face recognition, face detection, PCA, biometric.


This article describes all the steps, methodologies and algorithms used for the development of a biometric system based on a hybrid facial identification technique. The technique used is based on two different algorithms of face identification, working in parallel. One uses facial features localization, in order to compute a vector with 20 characteristics, extracted from an initial image. The other uses a much known holistic process usually called principal components analysis (PCA), in order to represent an image on a reduced dimension space. Due to this dichotomy we consider this a hybrid system. We use these two techniques on an independent way until the final decision. This biometric system starts when it receives a facial image of an individual, and the purpose is to identify it from an image database. Resuming, this article explains all the steps followed to build this system, as well as the theoretical principals that support them.

Important Links:

Go Back