On matching digital face images against scanned passport photos

Abstract

In this paper the problem of matching high-resolution digital face images against low-resolution passport photos scanned from the original document is studied. The challenges involved in such a problem are quite different from those encountered by classical facial matching systems described in the literature. These challenges can be person-related, document-related or scanning device-related. Thus, the matching performance of face recognition algorithms can be significantly impacted when tested on passport photos scanned by an external device. The purpose of this paper is to illustrate the complexity of the passport facial matching problem and to provide a preliminary solution to it. The contributions of this work are two-fold. Firstly, a database of 25 subjects is assembled and used to illustrate the challenges associated with passport facial matching. Secondly, a pre-processing scheme is proposed in order to denoise and eliminate watermark traces that may be present in scanned passport photos. The application of such a pre-processing scheme is observed to improve the performance of classical face matching systems. To the best of our knowledge, this is the first time that this problem is being investigated in the open literature in the context of international passports exhibiting a variety of facial image qualities and security marks.