Mauro Barni

Source Attribution of AI-Generated Images: a Principled Survey

Meiling Li, Benedetta Tondi, Pietro Bongini, Zhenxing Qian et autres

This record contains the preprint version of a survey manuscript on source attribution of AI-generated images. The manuscript reviews existing approaches for tracing synthetic images back to their generative sources, with a focus on passive attribution methods. It organizes the literature under …

cn, it (code pays fourni par la source)

Source Attribution of AI-Generated Images: a Principled Survey

Meiling Li, Benedetta Tondi, Pietro Bongini, Zhenxing Qian et autres

This record contains the preprint version of a survey manuscript on source attribution of AI-generated images. The manuscript reviews existing approaches for tracing synthetic images back to their generative sources, with a focus on passive attribution methods. It organizes the literature under …

cn, it (code pays fourni par la source)

Comparative Study of Adversarial Training and Randomized Smoothing for Robust AI-Generated Image Attribution

Kai Zeng, Niccolò Pancino, Nasrin Malekzadeh Goradel, Mauro Barni et autres

In this paper we explore two different approaches for designing AI-generated image attribution methods that are robust in adversarial settings, namely adversarial training (AT) and randomized smoothing (RS). While AT has been widely adopted in machine learning to improve the adversarial robustness …

it (code pays fourni par la source)

An efficient watermarking method for latent diffusion models via low-rank adaptation and dynamic loss weighting

Dongdong Lin, Yue Li, Benedetta Tondi, Bin Li et autres

cn, it (code pays fourni par la source)

Source Attribution of AI-Generated Images in Data Scarcity Conditions via Training-Free Resynthesis

Pietro Bongini, M C Li, Andrea Costanzo, Benedetta Tondi et autres

Attributing synthetic images to the source that generated them is a difficult problem, particularly in data-scarcity conditions requiring the adoption of few-shot or zero-shot learning strategies. In this paper, we tackle this problem by introducing a training-free attribution method based on image …

it, cn (code pays fourni par la source)

VocaLock: Watermark-Based Detection of Zero-Shot Voice Conversion Manipulation and Timbre Attribution

Yunming Zhang, Dengpan Ye, Benedetta Tondi, Mauro Barni

cn, it (code pays fourni par la source)

WILD: a new in-the-Wild Image Linkage Dataset for synthetic image attribution

Pietro Bongini, Sara Mandelli, Andrea Montibeller, Mirko Casu et autres

Synthetic image source attribution is an open challenge, with an increasing number of image generators being released yearly. The complexity and the sheer number of available generative techniques, as well as the scarcity of high-quality open source datasets of diverse nature for …

it (code pays fourni par la source)

Print and Scan Simulation for Adversarial Attacks on Printed Images

Nischay Purnekar, Benedetta Tondi, Jana Dittmann, Mauro Barni

Predictive AI with deep learning is vulnerable to adversarial examples—subtle, human-imperceptible modifications that can induce classification errors or evade detection. While most research targets digital adversarial attacks, many real-world applications require attacks to function in the physical domain. Physical adversarial examples must …

it, de (code pays fourni par la source)

Beyond Pixels: Sentiment-Based Assessment of Image Inpainting Techniques with Application to Forgery Removal

Marco Blanchini, Giovanna Maria Dimitri, Mauro Barni

Inpainting is an image processing technique used to remove alterations and restore the original content of a modified image. The effectiveness of inpainting can be evaluated through various methods, increasingly based on criteria related to human perception. In this work, we propose …

it (code pays fourni par la source)

WILD: a new in-the-Wild Image Linkage Dataset for synthetic image attribution

Pietro Bongini, Sara Mandelli, Andrea Montibeller, Mirko Casu et autres

Synthetic image source attribution is an open challenge, with an increasing number of image generators being released yearly. The complexity and the sheer number of available generative techniques, as well as the scarcity of high-quality open source datasets of diverse nature for …

Deepfake Media Forensics: Status and Future Challenges

Irene Amerini, Mauro Barni, Sebastiano Battiato, Paolo Bestagini et autres

The rise of AI-generated synthetic media, or deepfakes, has introduced unprecedented opportunities and challenges across various fields, including entertainment, cybersecurity, and digital communication. Using advanced frameworks such as Generative Adversarial Networks (GANs) and Diffusion Models (DMs), deepfakes are capable of producing highly …

it (code pays fourni par la source)

BOSC: A Backdoor-Based Framework for Open Set Synthetic Image Attribution

Jun Wang, Benedetta Tondi, Mauro Barni

With the continuous progress of AI technology, new generative architectures continuously appear, thus driving the attention of researchers towards the development of synthetic image attribution methods capable of working in open-set scenarios. Existing approaches focus on extracting highly discriminative features for closed-set …

it (code pays fourni par la source)