Reliability of personal identification base on optical 3D measurement of a few facial landmarks
Emanuele Zappa, Paolo Mazzoleni · Procedia Computer Science · 2010
Technologies related to identity recognition have found widespread application in the last years; among these, face recognition is one of the most promising and probably the most studied. The main goal of this work is to verify if it is possible to obtain a reliable identification of people, starting from a few 3D measurements of the face shape. To this purpose, an algorithm of face recognition based only on the 3D localization of some significant points of the face, starting from a pair of stereoscopic photographs, was developed and applied on a 51 people database. The localization of these landmarks is carried out manually: the goal of the work is the evaluation of the reliability of a few-point based algorithm, not its automation. The obtained results show a recognition percentage grater than 90%, an encouraging outcome for a preliminary work.