Abstract

GRENOT CASTELLANO, Elaine et al. Ears Images Normalization for Biometric Recognition. EAC [online]. 2019, vol.40, n.3, pp. 28-38.  Epub Sep 08, 2019. ISSN 1815-5928.

Image alignment is a crucial step in recognition systems based on biometric images, since it facilitates image analysis and comparison. In particular, ear biometric images may present occlusion problems because of hair, earrings or hats. These occlusions decrease the ear recognition zone; as well as the variation of lighting in the capture process increases difficulty of extracting traits or fiducial points. Mutual information is a statistical technique based on the information provided by the distribution of the intensities in images. The proposed method based on this technique does not depend on the actual values of the pixels but on how they are distributed in the image, this allows you to deal with problems of occlusion, effects of lighting, and noise. In this paper, we will show the usefulness of mutual information between ear images for their alignment. To do this, several experiments were conducted comparing this method with others of a statistical nature, such as those in PCA and other based on fiducial points, such as warping. The evaluation of results is made through recognition sequences by representing the ear with functions.

Keywords : Ear; alignment; mutual information.

        · abstract in Spanish     · text in Spanish     · Spanish ( pdf )