Réalisation d’un système de détection des visages en utilisant la matrice PSSM –positionspecific- scoring-matrix

Abstract

Although the recognition of human faces is a difficult field because of the multitude of parameters that must be taken into account (variation in posture, lighting, hair style, wearing glasses, beards, moustaches, old age...), it is very important to be interested given the many fields of applications (verification of people, remote surveillance, man-machine interfaces...). So we have the contribution of this thesis consisted in developing a facial recognition technique based on the PSSM matrix (position specific scoring matrix), which takes as an input, not the entire facial image, but the corresponding code sequences of the characteristic areas of the face, and to demonstrate that this method gives good recognition rates.