Videos versus still images: Asymmetric sensor pattern noise comparison on mobile phones

Galdi, Chiara; Hartung, Frank; Dugelay, Jean-Luc
EI 2017, Electronic Imaging, January 29-February 2nd, 2017, Burlingame, USA

Nowadays, the most employed devices for recoding videos or capturing images are undoubtedly the smartphones. Our work investigates the application of source camera identification on mobile phones. We present a dataset entirely collected by mobile
phones. The dataset contains both still images and videos collected by 67 different smartphones. Part of the images consists in photos of uniform backgrounds, especially collected for the computation of the RSPN. Identifying the source camera given a video
is particularly challenging due to the strong video compression. The experiments reported in this paper, show the large variation in performance when testing an highly accurate technique on still images and videos.

Type:
Conférence
City:
Burlingame
Date:
2017-01-29
Department:
Sécurité numérique
Eurecom Ref:
5123
Copyright:
IS&T

PERMALINK : https://www.eurecom.fr/publication/5123