In the last years, deep learning has had a huge impact on an incredible number of applications and deep learning solutions have also been widely explored and used to address several research questions. Deep learning-based techniques are also being utilized to generate extremely realistic manipulated facial multimedia content. User-friendly and easy-to use tools have enabled even non-experts to create manipulated contents with little effort. The pervasiveness of fake generated contents, combined with the velocity with which information can be spread through the use of social media, has raised the serious problem of the misinformation. The rapid advances in deep learning have brought to the continuous development of new and complex solutions for the generation of fakes contents (images, video and text) but also for their detection, with countermeasures that are being continuously developed by Multimedia Forensic (MMF) researchers to prevent their diffusion. The aim of this session is to present recent results and advances regarding deep learning methods for the generation and detection of fake media.

The scope of this session comprehends, but it is not limited to:

·      Deep Learning for Image and Video Forensics

·      Generation and Detection of Deepfakes

·      Image/Video Manipulation Detection

·      Audio Forensics

·      Counter-Forensics and Adversarial Attacks

·      Analysis of Text/Image inconsistencies

·      Detection of Manipulation Intent in multimedia (text, image, video, speech, etc)

KEY DATES AND DEADLINES

· Paper submission deadline: 12 June 2023  Extended!! 20th of June 2023

· Author notification: 12 July 2023

· Camera-ready submission: 18 July 2023

· All deadlines expire at 23:59 AoE (Anywhere on Earth, UTC - 12)

· Conference: 18-22 September, 2023

· Workshop Date: 22 September, 2023

Works will be included in a post-conference proceedings volume edited by Springer.