Open research “TrueBees: Digital forensics and blockchain technology against deceptive deep fake imaging in social media”

With prof. Giulia Boato and Daniele Miorandi

Visual contents are becoming more and more popular on social networks, with more than 3,2 billion images and 720000 hours of videos shared every day. It is therefore not surprising that there has been a long-standing interest by malicious users/organizations in manipulating visual contents and using them for diffusing unreliable information and fake news. While tampering with images has traditionally been a lengthy and labour-intensive process, recent advancements in artificial intelligence have made available imaging tools capable of automatically generating highly realistic fakes. On top of that, research studies have shown that human performance (even of experts) in distinguishing real pictures from synthetic ones is alarmingly poor. This has given rise to the phenomenon of deep fakes, which has threatened the trustworthiness of visual media from online sources, in particular from social media.

During this open research session we will present research activities carried out by the Media Lab @UNITN and the deep-tech SME U-Hopper in the TrueBees project which aims at restoring trust in visual content shared by users on social networks by combining innovative digital forensics tools with blockchain technology.

Registrazione

Mandatory registration for attendance in person. Fill out the online registration form and choose the Open Research you are interested in.

It is also possible to follow the seminar via Webinar on Zoom at
https://unitn.zoom.us/j/89357062412?pwd=ZFl4S2Y0QWE1SXdRbEQ5TEFOdmQyQT09
Passcode: 554910

About the speakers

Giulia Boato is Associate Professor at DISI. She is currently co-PI of the PRIN project “PREMIER – Preserving media trustworthiness in the artificial intelligence era” (2020-2023) and of the DARPA AI Explorations Reverse Engineering Deceptions project “UNCHAINED – Uncovering media manipulation chains through container and content detectable traces” (2020-2022). She has chaired various conferences in the field of multimedia forensics and is author of more than 140 papers with H-index 29

Daniele Miorandi is CEO of U-Hopper. Daniele holds a PhD from University of Padova in telecommunications engineering. He is recognised for his research works in networking, IoT and big data (180+ publications, 8,500+ citations, listed by Stanford in the top 100k scientists worldwide in both 2019 and 2020). 5 patents awarded, 3 companies co-founded. He coordinated 3 EU cooperative R&D&I projects (BIONETS, Social-IST, SUPERHUB) and acted as co-PI/site PI in more than 30