Thèse soutenue

Reconnaissance biométrique par fusion multimodale du visage et de l'iris

FR  |  
EN
Auteur / Autrice : Nicolas Morizet
Direction : Laurence Likforman-SulemAmara Amara
Type : Thèse de doctorat
Discipline(s) : Signal et images
Date : Soutenance en 2009
Etablissement(s) : Paris, ENST

Mots clés

FR

Mots clés contrôlés

Résumé

FR  |  
EN

Biometrics refers to automatic recognition of individuals based on their physiological and/or behavioral characteristics. Unimodal biometric systems allow person recognition based on a single source of biometric information but cannot guaranty a perfect identification. Such systems are sensitive to noisy sensor data, non-universality and lack of individuality of the chosen biometric trait, and susceptible to spoof attacks. Most of those problems can be alleviated by using multimodal biometric systems that combine several biometric signatures. In this thesis, we address several important issues related to multimodal biometrics. First, after describing a state of the art in multimodal fusion, we establish the link between the brain processing and some basic face recognition algorithms. Then, we underline the use of wavelets in various levels of the multimodal biometric system. Lastly, the exploration of new fusion techniques of biometric signatures deriving from face (friendly and non-invasive) and iris (one of the most accurate biometrics) modalities and large-scale statistical analyses on match-scores deriving from both modalities have led to a novel adaptive fusion method combining wavelets andstatistical moments.