Digital, Semantic and Physical Analysis of Media Integrity

Can you trust what you see?

In the age of Photoshop and social media, a picture may still be worth a thousand words, but not all those words are true all the time. The goal of our DiSPARITY research effort is to create techniques that can assess the truthfulness, or integrity, of an image or a video clip—thereby allowing users to determine whether and how much to rely upon the specific visual information at hand.  

Our research characterizes visual and semantic signatures of manipulated digital media (images and videos) by considering pixel-level attributes, by analyzing the physics of the scene, and by investigating the semantics of the image in the context of other related images and videos. We are also characterizing the genealogy of a manipulated image given a world data set. Key challenges to achieving these goals include the wide variety of acquisition devices, the increasing sophistication of manipulation tools and techniques (including rapid advances in computer graphics technology), and the sheer volume of data to be analyzed.

DiSPARITY is sponsored by DARPA’s MediFor Program. Dr. David Doermann is the Program Manager.

Research Team @ USC Collaborators
Prem Natarajan (PI)
Wael AbdAlmageed (co-PI)
Ram Nevatia (co-PI) [IRIS]
Huaigu Cao
Yue (Rex) Wu
Stephen Rawls
Ekraam Sabir (PhD student)
Zekun Li (PhD student)
Ayush Jaiswal (PhD student)
Luisa Verdoliva (University of Naples)
Christian Riess (University of Eriangen-Nuremberg)