GLFF: Global and Local Feature Fusion for AI-synthesized Image Detection

TL;DR

The paper introduces GLFF, a novel framework that combines global and local features for improved detection of AI-synthesized images, especially in challenging real-world scenarios with post-processing and new generation models.

Contribution

It proposes a multi-scale feature fusion approach and creates a new challenging dataset to enhance AI-synthesized image detection in real-world conditions.

Findings

GLFF outperforms existing methods on the new DF3 dataset.

The method maintains high accuracy across multiple open-source datasets.

Fusion of global and local features improves detection robustness.

Abstract

With the rapid development of deep generative models (such as Generative Adversarial Networks and Diffusion models), AI-synthesized images are now of such high quality that humans can hardly distinguish them from pristine ones. Although existing detection methods have shown high performance in specific evaluation settings, e.g., on images from seen models or on images without real-world post-processing, they tend to suffer serious performance degradation in real-world scenarios where testing images can be generated by more powerful generation models or combined with various post-processing operations. To address this issue, we propose a Global and Local Feature Fusion (GLFF) framework to learn rich and discriminative representations by combining multi-scale global features from the whole image with refined local features from informative patches for AI synthesized image detection. GLFF fuses information from two branches: the global branch to extract multi-scale semantic features and the local branch to select informative patches for detailed local artifacts extraction. Due to the lack of a synthesized image dataset simulating real-world applications for evaluation, we further create a challenging fake image dataset, named DeepFakeFaceForensics (DF 3 ), which contains 6 state-of-the-art generation models and a variety of post-processing techniques to approach the real-world scenarios. Experimental results demonstrate the superiority of our method to the state-of-the-art methods on the proposed DF 3 dataset and three other open-source datasets.