We present MERIT, an inference-time modular framework for multimodal misinformation detection that decomposes verification into four specialized modules: visual forensics, cross-modal alignment, retrieval-augmented claim verification, and calibrated judgment. On MMFakeBench, MERIT with GPT-4o-mini achieves 81.65% F1, outperforming all reported zero-shot baselines including GPT-4V with MMD-Agent (74.0% F1). A controlled same-model evaluation confirms gains stem from architectural design: MERIT achieves 6.14 points higher misinformation recall than MMD-Agent under identical model conditions, with per-class gains of +18.0 on visual distortion and +5.33 on textual distortion. Ablation studies reveal non-overlapping module specialization, where removing any module disproportionately degrades its target category while leaving others intact. Test set evaluation on 5,000 samples confirms generalization within 0.21 F1 points of validation results. The framework operates with any instruction-following vision-language model and produces citation-linked rationales for human review.

翻译：摘要：我们提出MERIT，一种推理时模块化框架，用于多模态虚假信息检测，将验证过程分解为四个专门模块：视觉取证、跨模态对齐、检索增强的声明验证以及校准判断。在MMFakeBench数据集上，使用GPT-4o-mini的MERIT实现了81.65%的F1分数，超过所有已报告的零样本基线方法，包括结合GPT-4V与MMD-Agent（74.0% F1）。受控同模型评估证实性能提升源于架构设计：在相同模型条件下，MERIT的虚假信息召回率比MMD-Agent高出6.14个百分点，其中视觉失真类别提升+18.0，文本失真类别提升+5.33。消融研究揭示了非重叠的模块专业化特性：移除任一模块会不成比例地降低其目标类别性能，而其他类别保持完整。在5000样本测试集上的评估验证了泛化能力，与验证结果相差仅0.21 F1分。该框架适用于任何遵循指令的视觉-语言模型，并生成可链接引用的推理过程供人工审查。