General and Domain-Specific Zero-shot Detection of Generated Images via Conditional Likelihood

TL;DR

CLIDE is a new zero-shot detection method that uses conditional likelihood approximation to identify AI-generated images across various domains, outperforming existing methods in general and domain-specific scenarios.

Contribution

Introduces CLIDE, a novel zero-shot detection approach based on conditional likelihood, enabling effective domain adaptation for detecting AI-generated images.

Findings

Achieves state-of-the-art performance on large-scale datasets.

Outperforms existing methods in domain-specific detection.

Demonstrates robustness and broad applicability.

Abstract

The rapid advancement of generative models, particularly diffusion-based methods, has significantly improved the realism of synthetic images. As new generative models continuously emerge, detecting generated images remains a critical challenge. While fully supervised, and few-shot methods have been proposed, maintaining an updated dataset is time-consuming and challenging. Consequently, zero-shot methods have gained increasing attention in recent years. We find that existing zero-shot methods often struggle to adapt to specific image domains, such as artistic images, limiting their real-world applicability. In this work, we introduce CLIDE, a novel zero-shot detection method based on conditional likelihood approximation. Our approach computes likelihoods conditioned on real images, enabling adaptation across diverse image domains. We extensively evaluate CLIDE, demonstrating SOTA performance on a large-scale general dataset and significantly outperform existing methods in domain-specific cases. These results demonstrate the robustness of our method and underscore the need of broad, domain-aware generalization for the AI-generated image detection task. Code is available at https://tinyurl.com/clide-detector.