Blog

October 2024

Two papers accepted at WIFS 2024

We are pleased to share that our latest contribution to image forensics research has led to the acceptance of two of our papers at the 16th IEEE International Workshop on Information Forensics and Security (WIFS/) 2024. 📚 Accepted Papers Tiny autoencoders are effective few-shot generative model detectors by Luca Bindini, Giulia Bertazzini, Daniele Baracchi, Dasara Shullani, Paolo Frasconi, and Alessandro Piva Beyond the Brush: Fully-automated Crafting of Realistic Inpainted Images by Giulia Bertazzini, Chiara Albisani Daniele Baracchi, Dasara Shullani, and Alessandro Piva Moreover, our first accepted paper is the outcome of a successful collaboration with Luca Bindini and Prof.

September 2024

Identifying and Verifying Synthetic Media in the Age of AI Manipulation

Dr. Daniele Baracchi and Dr. Dasara Shullani are happy to present the research topic Identifying and Verifying Synthetic Media in the Age of AI Manipulation. As topic editors, our aim is to bring together cutting-edge research that explores how to identify and verify AI-generated content. This includes finding ways to detect the subtle traces left by generative models and creating reliable tools for verifying the authenticity of multimedia content. Key dates to note:

May 2024

Two papers accepted on Pattern Recognition Letters

We’re happy to share that two of our papers have been accepted for publication in the Special Issue on Advances in Disinformation Detection and Media Forensics (A2DMF) in Pattern Recognition Letters. 📚 Accepted Papers Continual learning for adaptive social network identification by Simone Magistri, Daniele Baracchi, Dasara Shullani, Andrew D. Bagdanov, and Alessandro Piva Uncovering the authorship: Linking media content to social user profiles by Daniele Baracchi, Dasara Shullani, Massimo Iuliani, Damiano Giani, and Alessandro Piva The first accepted paper is the result of a fruitful collaboration with Simone Magistri and Prof.

April 2024

International Conference on Acoustics, Speech and Signal Processing

Dr. Daniele Baracchi and Dr. Dasara Shullani participated in the 49th International Conference on Acoustics, Speech, and Signal Processing in Seoul. They showcased their research during the poster session on Multimedia Forensics and Cybersecurity, as well as through an oral presentation in the Multimedia Forensics track. The conference featured a wealth of intriguing research ideas, with notable keynotes by Prof. Daniel D. Lee on Geometry and Latent Signal Representations in Machine Learning, and Prof.

March 2024

Paper accepted on IEEE Access

We are pleased to announce the publication of our latest paper in IEEE Access, a product of our strong collaboration with colleagues from the University of Trento as part of the UNCHAINED project. 📚 Accepted Paper Toward Open-World Multimedia Forensics Through Media Signature Encoding by Daniele Baracchi, Giulia Boato, Francesco de Natale, Massimo Iuliani, Andrea Montibeller, Cecilia Pasquini, Alessandro Piva, and Dasara Shullani About the authors Daniele Baracchi, Dasara Shullani, and Massimo Iuliani are postdoctoral researchers in the Department of Information Engineering at the University of Florence.

Machine Learning Summer School in Okinawa

Chiara Albisani and Giulia Bertazzini participated in the Machine Learning Summer School at the Okinawa Institute of Science and Technology. Throughout the program, they delved into advanced machine learning techniques presented by leading industry experts. Highlights included remarkable lessons on Large Language Models by Prof. Tatsunori Hashimoto and Natural Language Processing by Prof. Diyi Yang. They also had the opportunity to showcase their research through a poster session, engaging in stimulating discussions with peers and experts.

February 2024

Paper presentation at ICICT 2024

Yoshihisa Furushita eagerly attended the 9th International Congress on Information and Communication Technology in London, where he gave an oral presentation on his research, “Double Compression Detection of HEIF Images Using Coding Ghosts”. This was the inaugural presentation for Yoshihisa, who was able to use the occasion to meet researchers from all over the world. About the author Yoshihisa is a third-year doctoral candidate in the PhD Program in Information Engineering of the University of Florence.

January 2024

Appearance deceives: hunting down fake images and videos with multimedia forensics techniques

We are thrilled to announce a seminar as part of the upcoming Sarò Matricola event, dedicated to prospective freshmen of the School of Engineering. Dr. Daniele Baracchi will deliver a talk titled “L’apparenza inganna: a caccia di immagini e video falsi con tecniche di multimedia forensics” (“Appearance deceives: hunting down fake images and videos with multimedia forensics techniques”). The seminar will showcase the techniques of multimedia forensics that can be employed to discern the authenticity of images and videos.

Audio Forensics Seminar

We are happy to announce an upcoming seminar where our newest visiting posdoctoral researcher, Gül TahaoÄŸlu, will present on the topic of “Audio forensics”. Date: 2024-01-19 Time: 14:30 Location: DINFO Meeting Room (264) During this seminar, Gül will present the KTU Cyber Security Research Group and some of their recent reasearch endevors. Attendees can expect the following: Audio copy-move detection: via feature matching on MEL spectogram. Audio forgery localization: via a super-resolution spectogram and a keypoint-based clustering.

December 2023

Two papers accepted at ICASSP 2024

We are thrilled to announce that our dedication to multimedia forensics research has once again borne fruit! Two of our papers have been accepted at the International Conference on Acoustics, Speech, and Signal Processing (ICASSP/) 2024. 📚 Accepted Papers A codec-based approach for video life-cycle characterization in social networks by Giulia Bertazzini, Daniele Baracchi, Dasara Shullani, Massimo Iuliani, and Alessandro Piva Structure matters: analyzing videos via graph neural networks for social media platform attribution by Andrea Gemelli, Dasara Shullani, Daniele Baracchi, Simone Marinai, and Alessandro Piva We highlight that the first paper represents the inaugural publication of our talented first-year Ph.

Winter Refreshment for Junior Researchers

At LESC, we understand that research is not just about individual brilliance but also about the synergy created through collaboration. Therefore, we’re excited to invite you to a special winter refreshment event to celebrate our shared journey and strengthen our bonds as a research community. All PhD students and postdoctoral researchers of the department are warmly invited to join us for this special occasion. This event presents an opportunity for us to come together, regardless of our individual research interests, and connect as colleagues and friends.

November 2023

Image Quality Assessment with Deep Learning for medical images. Examples and opportunities

We are glad to share news of an upcoming seminar featuring Chiara Albisani, who will be presenting on the intriguing topic of “Image Quality Assessment with Deep Learning for medical images. Examples and opportunities”. Date: 2023-12-01 Time: 10:00 Location: Room 031, School of Engineering Chiara’s presentation will delve into the cutting-edge algorithms used for assessing image quality in medical imaging. By exploring the current state-of-the-art techniques, she will also shed light on promising research avenues within this field.

October 2023

An overview of Diffusion Models in Multimedia Forensics: Generating and Detecting Synthetic Images

We are happy to announce an upcoming seminar where our newest PhD student, Giulia Bertazzini, will present on the topic of “Diffusion Models in Multimedia Forensics: Generating and Detecting Images”. Date: 2023-10-27 Time: 10:00 Location: DINFO Meeting Room (264) During this seminar, Giulia will provide a comprehensive introduction to diffusion models and will explore current state-of-the-art techniques for detecting content produced by them. Attendees can expect the following: Introduction to Diffusion Models: Giulia will explain the underlying principles and functioning of diffusion models.

Welcome to Our XXXIX Cycle PhD Scholars

We are thrilled to extend a warm welcome to the newest additions to our research family! It is with great excitement that we introduce Chiara Albisani and Giulia Bertazzini, who have joined us on our journey towards academic excellence. Chiara brings with her expertise in medical image analysis using machine learning. She will receive guidance from Prof. Andrew D. Bagdanov and Prof. Fabrizio Argenti throughout her doctoral studies. Giulia’s research focuses on detecting synthetic images generated by diffusion models.

September 2023

FloreView: an image and video dataset for forensic analysis

We are pleased to announce the publication of our latest paper in IEEE Access. The paper, titled “FloreView: an image and video dataset for forensic analysis”, unveils a cutting-edge dataset for digital forensic analysis. The newly introduced dataset sets itself apart by offering a diverse collection of over 9000 media contents captured by 46 smartphones from 11 major brands. To learn more, visit the dedicated page for this paper.

Greetings to Our Visiting Scholars

With a warm welcome, we introduce our newest team members, Lian Zong and Gül Tahaoğlu. Each brings unique skills, fresh perspectives, and a dedication to advancing signal processing. Their presence promises to enhance our collaborative efforts and foster innovation within our lab. Lian is a PhD student from Nanjing University of Information Science and Technology, and her main research focus is on Urban Boundary Layer Processes and their Climatic Environmental Effects, and Application of LiDAR in Atmospheric Environment Monitoring.

April 2023

Multimedia content protection: from model-based to data-driven approaches

We are delighted to share that three members of our lab will give a seminar on “Multimedia content protection: from model-based to data-driven approaches” at the prestigious Computer Vision Center in Barcelona. Dr. Daniele Baracchi, Dr. Massimo Iuliani, and Dr. Dasara Shullani, along with Simone Magistri from the Global Optimization Laboratory and Prof. Andrew D. Bagdanov, will present the results of the research activity on multimedia content security carried out at the Department of Information Engineering at the University of Florence.

UNCHAINED project: LESC goes to Washington

We are excited to share that two representatives from LESC will soon be in Washington to showcase the advancements made in the UNCHAINED project at the headquarters of DARPA. The project, carried out in collaboration with the University of Trento, aims to develop methods capable of identifying the processing chain through which an image has been processed. The presentation will be delivered by Daniele Baracchi and Dasara Shullani from the University of Florence and by Andrea Montibeller from the University of Trento.

March 2023

One paper accepted at IWBF 2023

We are pleased to announce that one of our papers has been accepted at the International Workshop on Biometrics and Forensics, which will take place in Barcelona in April. The paper concerns the application of continual learning techniques to the forensic task of social network identification and is the result of a collaboration with Simone Magistri from the Global Optimization Laboratory and Prof. Andrew D. Bagdanov from the Department of Information Engineering.