Electronic identity documents are spreading worldwide and the digital identification procedures relying on them are increasingly important for national and international security. In this paper, we study the impact of lossy compression on face images within the electronic identity documents domain. The study is conducted following the enrolment process in accordance with the current regulations and their recent updates that push the compression ratio further. Several experiments were carried out using different datasets, each one reflecting a specific enrolment setting as suggested by ISO and ICAO incoming regulations. The results are controversial. Even though the commercial tool we used for our experiments reported only negligible variations after image compression, the well-known open-source tool we adopted showed a considerable worsening. Image size limited to the current specifications could waste the benefits derived from the enhancements in the enrolment process concerning the original photo size and the scanning/acquisition quality. As a cautionary remark, we thus believe it would be preferable to update the enrolment guidelines, specifying to increase the space reserved for the facial image.
Calderoni L., Magnani A. (2022). The impact of face image compression in future generation electronic identity documents. FORENSIC SCIENCE INTERNATIONAL. DIGITAL INVESTIGATION, 40, 1-8 [10.1016/j.fsidi.2022.301345].
The impact of face image compression in future generation electronic identity documents
Calderoni L.
Primo
;Magnani A.Secondo
2022
Abstract
Electronic identity documents are spreading worldwide and the digital identification procedures relying on them are increasingly important for national and international security. In this paper, we study the impact of lossy compression on face images within the electronic identity documents domain. The study is conducted following the enrolment process in accordance with the current regulations and their recent updates that push the compression ratio further. Several experiments were carried out using different datasets, each one reflecting a specific enrolment setting as suggested by ISO and ICAO incoming regulations. The results are controversial. Even though the commercial tool we used for our experiments reported only negligible variations after image compression, the well-known open-source tool we adopted showed a considerable worsening. Image size limited to the current specifications could waste the benefits derived from the enhancements in the enrolment process concerning the original photo size and the scanning/acquisition quality. As a cautionary remark, we thus believe it would be preferable to update the enrolment guidelines, specifying to increase the space reserved for the facial image.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.