Personal Identification Through 3D Biometric Measurements Based on Stereoscopic Image Pairs

Abstract

The possibility of identifying people starting form images grabbed by video surveillance systems is a key characteristic that transforms a deterrence device in an effective and useful investigation tool. A system that allows to identify people in a reliable way, permits to increase the security of citizens. This means that the surveillance system has to be properly designed and installed. Images coming from common surveillance systems, rarely show a sufficient quality, and it is often impossible to verify personal identity in a reliable way, even combining different techniques together. In this paper an innovative video surveillance system, is proposed; this system allows to grab couples of stereoscopic images for each person passing within the cameras field of view. Starting from these images it will be possible to extract the three-dimensional positions of some face features, allowing for a reliable personal identification