Milan, Lombardy, Italy, June 25, 2026 – The Politecnico di Milano is at the forefront of the fight against digital disinformation, announcing the successful completion of two European Union-funded projects, FF4ALL and FUN-Media, dedicated to the detection of deepfakes. These initiatives, spearheaded by the university’s Image and Sound Processing Lab (ISPL) within the Department of Electronics, Information and Bioengineering, represent a significant advancement in verifying digital content and protecting users from manipulated media.
Combating Digital Disinformation: A Collaborative Effort
The research, coordinated by Professors Stefano Tubaro and Paolo Bestagini, with key contributions from researchers Sara Mandelli and Luca Comanducci, focused on developing robust technologies to identify synthetic images, videos, and audio. This effort is crucial in an era where generative AI allows for the creation of increasingly realistic manipulated content, posing substantial risks of disinformation, manipulation, and social engineering.
FF4ALL: Tackling Visual Deepfakes
The FF4ALL project specifically addressed the analysis of techniques used to generate and distribute synthetic images and videos. The ISPL team developed innovative tools for detecting artificial faces by combining 3D information and structural features. This approach significantly improves the robustness of detection models, even when faced with compression and post-processing techniques commonly used to obscure manipulation. A notable outcome of this project is the creation of the WILD dataset, a comprehensive collection of images generated by state-of-the-art deepfake models, which will serve as a valuable resource for future research and development in this field.
FUN-Media: Addressing the Threat of Audio Deepfakes
Complementing the visual focus of FF4ALL, the FUN-Media project delved into the detection of audio deepfakes, an emerging and less explored threat. Researchers developed novel architectures based on “Mixture of Experts” models, designed to enhance detection performance in complex and continuously evolving scenarios. Further work within FUN-Media concentrated on anomaly-based detection and attribution techniques, aiming to identify the specific generative technology responsible for creating synthetic audio content. “Alongside detection, the project also addressed the issue of attribution, namely the identification of the generative technology responsible for creating an audio content,” stated Paolo Bestagini, professor at the Department of Electronics, Information and Bioengineering.
A Step Forward in Digital Content Verification
The successful completion of these two projects marks a significant stride towards creating a more secure digital environment. By providing reliable tools for detecting and attributing deepfakes, the Politecnico di Milano is contributing to the global effort to counter the spread of false information and safeguard the integrity of digital media. These advancements are essential for maintaining trust in digital communication and protecting individuals and society from the adverse effects of manipulated content.
The findings from FF4ALL and FUN-Media will undoubtedly influence future research and development in the field of digital forensics and content verification, further solidifying the Politecnico di Milano’s position as a leader in innovative technological solutions for societal challenges.
Source: https://www.openinnovation.regione.lombardia.it/it/news/news/9323/politecnico-di-milano-tackles-deepfakes
