UGD-IML: A Unified Generative Diffusion-based Framework for Constrained and Unconstrained Image Manipulation Localization

TL;DR

This paper introduces UGD-IML, a diffusion-based generative framework that unifies image manipulation localization and constrained localization, reducing data dependency and improving performance in forgery detection.

Contribution

The work presents the first unified diffusion model for both IML and CIML, enabling effective localization with limited data and seamless task switching without extra training.

Findings

Outperforms state-of-the-art methods by 9.66% and 4.36% in F1 scores for IML and CIML.

Effective in limited data scenarios due to generative diffusion modeling.

Excels in uncertainty estimation, visualization, and robustness.

Abstract

In the digital age, advanced image editing tools pose a serious threat to the integrity of visual content, making image forgery detection and localization a key research focus. Most existing Image Manipulation Localization (IML) methods rely on discriminative learning and require large, high-quality annotated datasets. However, current datasets lack sufficient scale and diversity, limiting model performance in real-world scenarios. To overcome this, recent studies have explored Constrained IML (CIML), which generates pixel-level annotations through algorithmic supervision. However, existing CIML approaches often depend on complex multi-stage pipelines, making the annotation process inefficient. In this work, we propose a novel generative framework based on diffusion models, named UGD-IML, which for the first time unifies both IML and CIML tasks within a single framework. By learning the underlying data distribution, generative diffusion models inherently reduce the reliance on large-scale labeled datasets, allowing our approach to perform effectively even under limited data conditions. In addition, by leveraging a class embedding mechanism and a parameter-sharing design, our model seamlessly switches between IML and CIML modes without extra components or training overhead. Furthermore, the end-to-end design enables our model to avoid cumbersome steps in the data annotation process. Extensive experimental results on multiple datasets demonstrate that UGD-IML outperforms the SOTA methods by an average of 9.66 and 4.36 in terms of F1 metrics for IML and CIML tasks, respectively. Moreover, the proposed method also excels in uncertainty estimation, visualization and robustness.