The paper “Camera Attribution Forensic Analyzer in the Encrypted Domain” authored by atlanTTic researchers Alberto Pedrouzo-Ulloa, Miguel Masciopinto and Fernando Pérez-González, in cooperation with Juan R. Troncoso-Pastoriza, from EPFL (Switzerland) has received the Best Paper Award at the 2018 IEEE Workshop on Information Forensics and Security (WIFS) that took place in Hong Kong in December.
This paper proposes for the first time a solution that allows to establish whether an image was taken from a certain camera by preserving the privacy of the contents. The idea solves the problem of processing sensitive contents by third parties (for example, with Cloud services) with no privacy violations, because at all times the signals involved are encrypted. For instance, this would allow law enforcement agencies determine in child porrnography cases whether a picture was taken with the suspect’s camera by outsourcing the computations to the Cloud web services. The possibility of using invisible camera fingerprints to solve the so-called “camera attribution problem” was already known, but this is the first time the challenging problem of doing it with encrypted signals is formulated and solved.