I am a researcher at the Image and Visual Representation Lab (IVRL) at the École Polytechnique Fédérale de Lausanne, in Switzerland.

Prior to that, I worked at École Normale Supérieure Paris-Saclay, Université Paris-Saclay, Centre Borelli, CNRS, in Gif-sur-Yvette, France, where I obtained my PhD in applied mathematics in 2021.

My main interests currently are image processing, computer vision, machine learning, multimedia forensics and reproducible research.

Most notably, I invented the concept of Positional Learning, a technique to reveal, mimic and analyse underlying frequency components in images. This technique is seeing great uses in forensics, for forgery detection and AI-generated image detection.

Some of my image forensics methods are being used in the InVID-WeVerify verification plugin for fact-checkers, called by the Poynter Institute (home of the International Fact-Checking Network) “One of the most powerful tools for spotting misinformation online”.

I am heavily invested in the IPOL (Image Processing On Line) journal and demo system for open science and reproducible research. In particular, I am a cofounder and the main organizer of the IPOL MLBriefs workshop and hackhathon since its creation in April 2022. Four editions have already taken place, the last one was held in May, 2024.

Between 2022 and 2024, I was involved in the coordination and development of the BrevetAI platform, to offer a learning-by-doing training on artificial intelligence and disseminate knowledge about AI to the public. This platform is being developed as part of the SaclAI-School training program of the Université Paris-Saclay, piloted by the DATAIA institute.

You can contact me at firstname.lastname@epfl.ch (My first name is Quentin, my last name is Bammey).

Positional Learning

Positional Learning

I invented POsitional LeARning (Polar), a novel training scheme for CNNs that implicitly reveals underlying spatial or spectral information and can highlight its inconsistencies. Positional learning has already been applied successfully to AI-generated image detection, and mosaic inconsistency detection, for image forensics. Continue reading Positional Learning