Abstract:

Rapid progress in machine learning, computer vision and graphics leads to successive democratization of media manipulation capabilities. While convincing photo and video manipulation used to require substantial time and skill, modern editors bring (semi-) automated tools that can be used by everyone. Some of the most recent examples include manipulation of human faces, e.g., by their replacement or semantic manipulation (expression, age, etc.). At the same time, dissemination of fake news and misinformation campaigns are picking up speed which challenges trust in the society. This talk presents techniques from digital media forensics that could be used for detection of manipulated media. We will explain the fundamental principles, discuss the most successful techniques, the biggest challenges and some of the most promising research directions. We also show connections to other fields like information hiding, computational photography, lossy compression and machine learning security.

BIO:

Paweł Korus is a research scientist working on protection and verification of digital media integrity. He received his Ph.D. degree in telecommunications from AGH University of Science and Technology, Krakow, Poland (2013) and did his postdoctoral research at the College of Information Engineering, Shenzhen University, China (2015-2017). Currently, he is a research assistant professor at the Center for Cybersecurity, New York University, USA. His research interests include various aspects of multimedia security, image processing, low-level vision and machine learning with particular focus on content authentication techniques for digital photographs. Dr Korus currently serves the research community as an elected member of the IEEE Information Forensics & Security Technical Committee. His work was supported by several research projects in the EU, China and the US and by a scholarship for outstanding young scientists from the Polish Ministry of Science and Higher Education. He is currently developing an open source library for modeling and optimization of next-generation image acquisition and distribution pipelines (https://github.com/pkorus/neural-imaging).

Homepage:  https://pkorus.pl and https://github.com/pkorus/

Sponsored in part by:   Facebook Reality Labs Pittsburgh