Person Recognition Beyond The Visible Spectrum: Combining Body Shape and Texture from mmW images

Biometrics have been tradittionally based on images acquired in the visible spectrum. In this paper, we will go first into details regarding the regions beyond the visible spectrum that have been already explored in the literature for biometrics to overcome some of the limitations found in the visible region. Later, we will introduce millimeter imaging as a new region of the spectrum that has also potential in biometrics. To this aim, we first consider shape and texture information individually for person recognition. Later, we compare them and study to what extent the joint use of shape and texture can provide further improvements. Results suggest that both sources of information can complement each other, reaching verification results of 1.5% EER. This result motivates us to think that in the future, person recognition can be integrated within the millimeter screening scanners already deployed in airports, and enhance this way security.