DA-HFNet: Progressive Fine-Grained Forgery Image Detection and Localization Based on Dual Attention

TL;DR

This work tackles the challenge of detecting and localizing forged images from AIGC by introducing DA-HFNet, a progressive, multi-feature framework. It fuses RGB, noise, and frequency information through a dual-attention mechanism and leverages an HRNet-based multi-scale interaction to capture forgery cues at varying resolutions. A progressive detection-localization module propagates priors across hierarchical levels, aided by an edge-aware loss to refine boundaries. The authors also construct the DA-HFNet forged-image dataset with text- and image-guided forgery from GANs and diffusion models, and demonstrate significant gains over state-of-the-art methods in both detection and localization, including cross-dataset validation. Overall, the approach advances robust forgery analysis for high-quality AIGC images, with practical implications for digital forensics and content verification.

Abstract

The increasing difficulty in accurately detecting forged images generated by AIGC(Artificial Intelligence Generative Content) poses many risks, necessitating the development of effective methods to identify and further locate forged areas. In this paper, to facilitate research efforts, we construct a DA-HFNet forged image dataset guided by text or image-assisted GAN and Diffusion model. Our goal is to utilize a hierarchical progressive network to capture forged artifacts at different scales for detection and localization. Specifically, it relies on a dual-attention mechanism to adaptively fuse multi-modal image features in depth, followed by a multi-branch interaction network to thoroughly interact image features at different scales and improve detector performance by leveraging dependencies between layers. Additionally, we extract more sensitive noise fingerprints to obtain more prominent forged artifact features in the forged areas. Extensive experiments validate the effectiveness of our approach, demonstrating significant performance improvements compared to state-of-the-art methods for forged image detection and localization.The code and dataset will be released in the future.

Figures (4)

• Figure 1: DA-HFNet Synthetic Image Dataset Examples.It is a dataset that is based on both GAN and diffusion models, encompassing both completely fabricated and partially edited images.
• Figure 2: DA-HFNet Model Structure.
• Figure 3: Dual Attention Feature Fusion Module.
• Figure 4: Comparison of Synthetic Region Localization Results.