ML Security PapersOmniDFA: A Unified Framework for Open Set Synthesis Image Detection and Few-Shot Attribution
Shiyu Wu 1,2,3, Shuyan Li 4, Jing Li 5, Jing Liu 1,3, Yequan Wang 2
Published on arXiv
2509.25682
Output Integrity Attack
OWASP ML Top 10 — ML09
Key Finding
OmniDFA achieves state-of-the-art generalization on AIGI detection while enabling few-shot attribution of unseen generative models using a 1.17M-image benchmark spanning 45 distinct generators.
OmniDFA
Novel technique introduced
AI-generated image (AIGI) detection and source model attribution remain central challenges in combating deepfake abuses, primarily due to the structural diversity of generative models. Current detection methods are prone to overfitting specific forgery traits, whereas source attribution offers a robust alternative through fine-grained feature discrimination. However, synthetic image attribution remains constrained by the scarcity of large-scale, well-categorized synthetic datasets, limiting its practicality and compatibility with detection systems. In this work, we propose a new paradigm for image attribution called open-set, few-shot source identification. This paradigm is designed to reliably identify unseen generators using only limited samples, making it highly suitable for real-world application. To this end, we introduce OmniDFA (Omni Detector and Few-shot Attributor), a novel framework for AIGI that not only assesses the authenticity of images, but also determines the synthesis origins in a few-shot manner. To facilitate this work, we construct OmniFake, a large class-aware synthetic image dataset that curates $1.17$ M images from $45$ distinct generative models, substantially enriching the foundational resources for research on both AIGI detection and attribution. Experiments demonstrate that OmniDFA exhibits excellent capability in open-set attribution and achieves state-of-the-art generalization performance on AIGI detection. Our dataset and code will be made available.
Key Contributions
- OmniDFA: a unified framework that combines open-set AI-generated image detection with few-shot source model attribution in a single system
- OmniFake: a large-scale dataset of 1.17M synthetic images curated from 45 distinct generative models to support both detection and attribution research
- Open-set few-shot source identification paradigm enabling attribution of unseen generators using only limited samples
🛡️ Threat Analysis
OmniDFA directly addresses AI-generated image detection and source model attribution — both are core output integrity concerns. OmniFake (1.17M images from 45 generators) and the open-set few-shot attribution paradigm are specifically designed to determine content provenance and combat deepfake abuse.