Digital Visual Media Forensics:
Lessons in Anti–Forensics and Counter Anti–Forensics
The tendency to believe what we see is a biological predilection, which is why photographs and videos are among the most influential and persuasive forms of evidence we have at our disposal today. As evidence (in both a general and legal sense of the word), photographs and videos are epistemologically unique. They act as silent eyewitnesses and because they are often perceived as a representational, or perhaps even an ontological, equivalent of what they depict, their testimony often becomes incontestable. But for us to use the evidence of photographs and videos safely, let alone effectually, it is essential ― as in the case of all other kinds of evidence ― to be aware of their flaws, the most consequential amongst which is their susceptibility to conscious semantic manipulations.
Photographs and videos are highly vulnerable to manipulation, but this vulnerability does little to preclude them from being treated as evidence. What it does do, however, is engender the need to devise powerful tools and techniques that can help verify the credibility of this evidence. This is made possible by a special branch of digital forensics, known as Digital Visual Media Forensics (DVMF). DVMF deals with the recovery and repair, scientific examination and investigation, comparison, and in–depth evaluation and interpretation of digital visual evidence, so as to ascertain its integrity and legitimacy, in mostly (civil or criminal) legal matters. DVMF is a multifaceted discipline, and among all the operations it entails, detection of conscious semantic manipulations (i.e., tamperings or forgeries) is perhaps the most explored. Over two decades of research in this field has resulted in a considerable body of specialized scientific procedures aimed at detecting the signs of all potential visual manipulations that can affect the integrity of a digital image or video. Unfortunately, the public availability of almost all of this research provides forgers with plenty of insight to create improved tampering schemes (called anti–forensic schemes), which are even more immune to detection than their original counterparts. Anti–forensic tampering schemes then require forensic analysts and researchers to develop counter anti–forensic strategies, which are specifically aimed at detecting anti–forensically created forgeries. Developments in the domain of counter anti–forensics are especially significant because without them, any advancement in the tamper detection domain is futile.
This course is dedicated to the comprehensive examination of all aspects of the domains of image and video anti–forensics and counter anti–forensics. The participants will gain a deep understanding of these highly consequential research domains, beginning with an introduction to the field of digital image and video forensics, the known kinds of image and video forgeries, the existing literature pertaining to the developments in the domain of forgery detection, and the various issues and challenges that complicate the task of forensic analysis of digital images and videos and impose restrictions on the extent of success of contemporary forensic solutions in heterogeneous real–world situations. The participants will learn about the process of forensic analysis of video evidence, including integrity verification, source authentication, and content authentication. The participants will also learn about the origin and history of deepfakes and take some time to examine some famous fake videos posted on YouTube. This will be an enjoyable exercise for the participants and will also enable them be more critical of the content we have become so accustomed to consume and inherently believe in our daily lives. This discussion will be followed by a detailed analysis of existing image and video anti–forensic and counter anti–forensic strategies, and will be concluded with an examination of the current research gaps, open issue and future research avenues in this domain. Altogether, this course will enable the participants to establish a strong foundation for personal growth and the eventual development of a meaningful career in the tremendously exciting domain of digital visual media forensics.