Towards Refining ID Cards Presentation Attack Detection Systems using Face Quality Index

Remote identity verification systems which rely on official documents, such as government-issued ID cards, require the use of Presentation Attack Detection (PAD) systems in order to combat impostor attacks. These attacks might use composition or splicing techniques, a printed ID card, or an image of the document displayed on a screen, altered prior to its presentation to the verification system. Based on recent efforts, Face Image Quality Assessment may help PAD systems in identifying ID card image tampering by utilising the photo ID section of the document, where a low face image quality could be related to printed, screen, or composite spoofed ID cards. This work presents an early approach towards complementing ID cards PAD systems by employing MagFace, a quality-aware face recognition method, in conjunction with a Chilean ID card PAD system, analysing the influence of the quality index of photo IDs on the system and its effectiveness when a prior rejection of low quality ID card images is made.