Sift Based Recognition of Finger Knuckle Print

Baptiste Hemery (1) , Romain Giot (1) , Christophe Rosenberger (1)

1 Equipe SAFE - Laboratoire GREYC - UMR6072

In this paper, we present a new unimodal method for finger knuckle print authentication. Previous works on this recent biometric modality mainly use image transform such as Radon, Gabor or Fourier. The proposed method is based on the computation of SIFT descriptors on the previously enhanced images of finger knuckle. Experiments on the recent PolyU FKP database show interesting results.

Domaines: Cryptographie et sécurité [cs.CR]

Baptiste Hemery, Romain Giot, Christophe Rosenberger. Sift Based Recognition of Finger Knuckle Print. The third Norsk Information security conference (NISK), 2010, Norway. pp.45-56. ⟨hal-00995114⟩