Attribution as Retrieval: Model-Agnostic AI-Generated Image Attribution

TL;DR

This paper introduces LIDA, a model-agnostic framework for attributing AI-generated images by treating it as an instance retrieval task, overcoming limitations of traditional classification methods.

Contribution

It proposes a novel, scalable, and model-agnostic approach for AI-generated image attribution using low-bit fingerprints and retrieval techniques.

Findings

Achieves state-of-the-art results in deepfake detection and attribution

Effective under zero- and few-shot learning scenarios

Demonstrates scalability to unseen generative models

Abstract

With the rapid advancement of AIGC technologies, image forensics will encounter unprecedented challenges. Traditional methods are incapable of dealing with increasingly realistic images generated by rapidly evolving image generation techniques. To facilitate the identification of AI-generated images and the attribution of their source models, generative image watermarking and AI-generated image attribution have emerged as key research focuses in recent years. However, existing methods are model-dependent, requiring access to the generative models and lacking generality and scalability to new and unseen generators. To address these limitations, this work presents a new paradigm for AI-generated image attribution by formulating it as an instance retrieval problem instead of a conventional image classification problem. We propose an efficient model-agnostic framework, called Low-bIt-plane-based Deepfake Attribution (LIDA). The input to LIDA is produced by Low-Bit Fingerprint Generation module, while the training involves Unsupervised Pre-Training followed by subsequent Few-Shot Attribution Adaptation. Comprehensive experiments demonstrate that LIDA achieves state-of-the-art performance for both Deepfake detection and image attribution under zero- and few-shot settings. The code is at https://github.com/hongsong-wang/LIDA