Research

Paper

AI LLM March 02, 2026

Process Over Outcome: Cultivating Forensic Reasoning for Generalizable Multimodal Manipulation Detection

Authors

Yuchen Zhang, Yaxiong Wang, Kecheng Han, Yujiao Wu, Lianwei Wu, Li Zhu, Zhedong Zheng

Abstract

Recent advances in generative AI have significantly enhanced the realism of multimodal media manipulation, thereby posing substantial challenges to manipulation detection. Existing manipulation detection and grounding approaches predominantly focus on manipulation type classification under result-oriented supervision, which not only lacks interpretability but also tends to overfit superficial artifacts. In this paper, we argue that generalizable detection requires incorporating explicit forensic reasoning, rather than merely classifying a limited set of manipulation types, which fails to generalize to unseen manipulation patterns. To this end, we propose REFORM, a reasoning-driven framework that shifts learning from outcome fitting to process modeling. REFORM adopts a three-stage curriculum that first induces forensic rationales, then aligns reasoning with final judgments, and finally refines logical consistency via reinforcement learning. To support this paradigm, we introduce ROM, a large-scale dataset with rich reasoning annotations. Extensive experiments show that REFORM establishes new state-of-the-art performance with superior generalization, achieving 81.52% ACC on ROM, 76.65% ACC on DGM4, and 74.9 F1 on MMFakeBench.

Metadata

arXiv ID: 2603.01993
Provider: ARXIV
Primary Category: cs.CV
Published: 2026-03-02
Fetched: 2026-03-03 04:34

Related papers

Raw Data (Debug)
{
  "raw_xml": "<entry>\n    <id>http://arxiv.org/abs/2603.01993v1</id>\n    <title>Process Over Outcome: Cultivating Forensic Reasoning for Generalizable Multimodal Manipulation Detection</title>\n    <updated>2026-03-02T15:45:33Z</updated>\n    <link href='https://arxiv.org/abs/2603.01993v1' rel='alternate' type='text/html'/>\n    <link href='https://arxiv.org/pdf/2603.01993v1' rel='related' title='pdf' type='application/pdf'/>\n    <summary>Recent advances in generative AI have significantly enhanced the realism of multimodal media manipulation, thereby posing substantial challenges to manipulation detection. Existing manipulation detection and grounding approaches predominantly focus on manipulation type classification under result-oriented supervision, which not only lacks interpretability but also tends to overfit superficial artifacts. In this paper, we argue that generalizable detection requires incorporating explicit forensic reasoning, rather than merely classifying a limited set of manipulation types, which fails to generalize to unseen manipulation patterns. To this end, we propose REFORM, a reasoning-driven framework that shifts learning from outcome fitting to process modeling. REFORM adopts a three-stage curriculum that first induces forensic rationales, then aligns reasoning with final judgments, and finally refines logical consistency via reinforcement learning. To support this paradigm, we introduce ROM, a large-scale dataset with rich reasoning annotations. Extensive experiments show that REFORM establishes new state-of-the-art performance with superior generalization, achieving 81.52% ACC on ROM, 76.65% ACC on DGM4, and 74.9 F1 on MMFakeBench.</summary>\n    <category scheme='http://arxiv.org/schemas/atom' term='cs.CV'/>\n    <published>2026-03-02T15:45:33Z</published>\n    <arxiv:primary_category term='cs.CV'/>\n    <author>\n      <name>Yuchen Zhang</name>\n    </author>\n    <author>\n      <name>Yaxiong Wang</name>\n    </author>\n    <author>\n      <name>Kecheng Han</name>\n    </author>\n    <author>\n      <name>Yujiao Wu</name>\n    </author>\n    <author>\n      <name>Lianwei Wu</name>\n    </author>\n    <author>\n      <name>Li Zhu</name>\n    </author>\n    <author>\n      <name>Zhedong Zheng</name>\n    </author>\n  </entry>"
}