Date and Time:
Tuesday, March 19, 11:00 – 13:00

This workshop is provided only face-to-face.

Description:

This workshop will provide hands-on deepfake making, as well as deepfake detection. We will provide new methods for making deepfakes, not only with face swaps, but also based on stable diffusion. The current deepfake software are improving rapidly, so detection is getting harder with the standard tools. We will cover classic ways of deepfake detection based on Photo Response Non Uniformity (PRNU) as well as double compression detection in video. Also we will discuss possibilities that videos made with deepfake software contain a hidden message or a watermark so detection gets easier. Also other methods such as the ENF (Electric Network Frequency) are discussed and with the rolling shutter effect a time of recording can be determined. In the hands on part deepfakes can be made with software on google co-lab. The detection can also be done with on line software that we will provide during the workshop. The use of AI to detect deepfakes is also presented.

Preparation details:

A laptop or at least a mobile phone with droid or iphone is needed though laptop with Windows or Unix and internet connection is preferred.

Workshop organiser:

Zeno Geradts is a senior forensic scientist at the Netherlands Forensic Institute of the Ministry of Security and Justice at the Forensic Digital Biometrics Traces departement. He is an expert witness in the area of forensic (video) image processing and biometrics such as manipulation detection on deepfakes. Within the team Forensic Big Data Analysis he works in research on artificial intelligence on text, images and video. He works within the European Project ASGARD on Forensic big data analysis. He was President of the American Academy of Forensic Science 2019-2020 and chairman of the ENFSI Forensic IT Working group. From September 1st 2014, he is full professor on Forensic Data Science by special appointment at the University of Amsterdam for 1 day a week. He has been active in forensic science since 1991 and has received several awards including the Distinguished Forensic Scientist award from ENFSI in 2012.

Merel de Leeuw den Bouter is a postdoctoral researcher in the AI4Forensics lab of the University of Amsterdam and the Netherlands Forensic Institute. Her research focuses on the forensic detection of deepfake videos using artificial intelligence. She has a PhD in applied mathematics from the Delft University of Technology.