This project will develop techniques to detect the use of steganography, which encompasses various methods of hiding messages in apparently ordinary images and videos. Because good steganography is designed to be invisible, the science of its detection, steganalysis, is difficult. This research will look for various signatures that might be left behind by the major classes of image and video steganography, principally by analyzing the structure of noise. This should lead to a novel class of steganography searchers based on image representations that depend on a quality factor, with the long-term goal being automated scanners that can rapidly find likely candidates amongst large numbers of images and videos. Slower and more specific methods can then decode the exact type of steganography used. The work will train students and introduce them to intelligence techniques, and further the collaboration between the University of Delaware and the U.S. Army Research Laboratory.
This award is supported jointly by the NSF and the Intelligence Community. The Approaches to Combat Terrorism Program in the Directorate for Mathematical and Physical Sciences supports new concepts in basic research and workforce development with the potential to contribute to national security.
