Visible to the public Threat Models and Games for Adversarial Multimedia Forensics

TitleThreat Models and Games for Adversarial Multimedia Forensics
Publication TypeConference Paper
Year of Publication2017
AuthorsBarni, Mauro, Tondi, Benedetta
Conference NameProceedings of the 2Nd International Workshop on Multimedia Forensics and Security
PublisherACM
Conference LocationNew York, NY, USA
ISBN Number978-1-4503-5034-1
Keywordsadversarial signal processing, game theoretic security, game theory, human factors, Metrics, multimedia forensics, pubcrawl, Scalability, simultaneous and sequential games, threat models for multimedia forensics
Abstract

We define a number of threat models to describe the goals, the available information and the actions characterising the behaviour of a possible attacker in multimedia forensic scenarios. We distinguish between an investigative scenario, wherein the forensic analysis is used to guide the investigative action and a use-in-court scenario, wherein forensic evidence must be defended during a lawsuit. We argue that the goals and actions of the attacker in these two cases are very different, thus exposing the forensic analyst to different challenges. Distinction is also made between model-based techniques and techniques based on machine learning, showing how in the latter case the necessity of defining a proper training set enriches the set of actions available to the attacker. By leveraging on the previous analysis, we then introduce some game-theoretic models to describe the interaction between the forensic analyst and the attacker in the investigative and use-in-court scenarios.

URLhttp://doi.acm.org/10.1145/3078897.3080533
DOI10.1145/3078897.3080533
Citation Keybarni_threat_2017