Search papers, labs, and topics across Lattice.
This paper introduces PromptForge-350k, a large-scale dataset for prompt-based AI image forgery localization, created using a novel automated mask annotation framework based on keypoint alignment and semantic similarity. To effectively leverage this dataset, they propose ICL-Net, a forgery localization network with a triple-stream backbone and intra-image contrastive learning to extract robust forensic features. Experiments show ICL-Net achieves state-of-the-art performance on PromptForge-350k and demonstrates strong robustness and generalization to unseen editing models.
AI-generated image forgery detection gets a major boost with PromptForge-350k, a dataset so large and well-annotated it pushes IoU scores 5% higher and generalizes to unseen models.
The rapid democratization of prompt-based AI image editing has recently exacerbated the risks associated with malicious content fabrication and misinformation. However, forgery localization methods targeting these emerging editing techniques remain significantly under-explored. To bridge this gap, we first introduce a fully automated mask annotating framework that leverages keypoint alignment and semantic space similarity to generate precise ground-truth masks for edited regions. Based on this framework, we construct PromptForge-350k, a large-scale forgery localization dataset covering four state-of-the-art prompt-based AI image editing models, thereby mitigating the data scarcity in this domain. Furthermore, we propose ICL-Net, an effective forgery localization network featuring a triple-stream backbone and intra-image contrastive learning. This design enables the model to capture highly robust and generalizable forensic features. Extensive experiments demonstrate that our method achieves an IoU of 62.5% on PromptForge-350k, outperforming SOTA methods by 5.1%. Additionally, it exhibits strong robustness against common degradations with an IoU drop of less than 1%, and shows promising generalization capabilities on unseen editing models, achieving an average IoU of 41.5%.
Tsinghua AI