Authors: M. Ferrara, A. Franco, D. Maltoni

C.d.L. Scienze dell’Informazione - Università di Bologna, via Sacchi 3, 47023 Cesena, Italy
DEIS - Viale Risorgimento, 2 - 40126 Bologna, Italy.
E-mail: {ferrara, franco, maltoni}@csr.unibo.it

1. Introduction

Face represents one of the most used biometric traits, for both computer automated and human assisted person identification. To allow interoperability among systems developed by different vendors and simplify the integration of biometric recognition in large-scale identification (e-passport, visas, etc.) a standard data format for digital face images is needed. In this context, the International Civil Aviation Organization (ICAO) started in 1980 a project focused on machine assisted biometric identity confirmation of persons. Initially three different biometric characteristics where identified for possible application in this context (face, fingerprint, iris), but finally face was selected as the most suited to the practicalities of travel document issuance, with fingerprint and/or iris available for choice by States for inclusion as complementary biometric technologies. Of course high quality, defect-free digital face images are needed to maximize both the human and computer assisted recognition accuracy. Starting from the ICAO work, in 2004 the International Standard Organization (ISO) defined a standard [3] for the digital face images to be used in the Machine Readable Travel Documents. The standard specifies a set of characteristics that the image has to comply, mainly related to the position of the face in the image and to the absence of defects (blurring, red eyes, face partially occluded by accessories, etc.) that would affect both the human and automatic recognition performance.