Performance characterisation of face recognition algorithms and their sensitivity to severe illumination changes

Abstract

This paper details the results of a face verification competition [2] held in conjunction with the Second International Conference on Biometric Authentication, The contest was held on the publically available XM2VTS database [4] according to a defined protocol [15]. The aim of the competition was to assess the advances made in face recognition since 2003 and to measure the sensitivity of the tested algorithms to severe changes in illumination conditions, In total, more than 10 algorithms submitted by three groups were compared1. The results show that the relative performance of some algorithms is dependent on training conditions (data, protocol) as well as environmental changes. © Springer-Verlag Berlin Heidelberg 2005.