MedGround: 通过验证的 grounding 数据弥合医学视觉-语言模型中的证据差距
MedGround: Bridging the Evidence Gap in Medical Vision-Language Models with Verified Grounding Data
AI总结：
MedGround通过生成高质量的医学指称 grounding 数据，提升视觉-语言模型在临床场景中的定位和泛化能力。
摘要：
视觉-语言模型（VLMs）可以生成令人信服的临床叙述，但经常难以将陈述视觉化地定位。我们提出这一限制源于高质量、大规模的临床指称定位对稀缺性。为此，我们引入了MedGround，一个自动化流程，将分割资源转换为高质量的医学指称 grounding 数据。利用专家掩膜作为空间锚点，MedGround 精确推导定位目标，提取形状和空间线索，并指导 VLMs 合成自然、临床 grounded 的查询，反映形态和位置。为了确保数据严谨性，多阶段验证系统整合了严格的格式检查、几何和医学先验规则以及基于图像的视觉判断，以过滤掉模糊或视觉上不支持的样本。最后，我们提出了MedGround-35K，一个新型多模态医学数据集。广泛的实验表明，使用MedGround-35K训练的VLMs在指称 grounding 性能、多对象语义歧义消除和对未见过的 grounding 设置的泛化能力方面均有所提高。这项工作突显了MedGround作为可扩展、数据驱动的方法，将医学语言与可验证的视觉证据联系起来。数据集和代码将在接受后公开发布。
英文摘要：
Vision-Language Models (VLMs) can generate convincing clinical narratives, yet frequently struggle to visually ground their statements. We posit that this limitation arises from the scarcity of high-quality, large-scale clinical referring-localization pairs. To address this, we introduce MedGround, an automated pipeline that transforms segmentation resources into high-quality medical referring grounding data. Leveraging expert masks as spatial anchors, MedGround precisely derives localization targets, extracts shape and spatial cues, and guides VLMs to synthesize natural, clinically grounded queries that reflect morphology and location. To ensure data rigor, a multi-stage verification system integrates strict formatting checks, geometry- and medical-prior rules, and image-based visual judging to filter out ambiguous or visually unsupported samples. Finally, we present MedGround-35K, a novel multimodal medical dataset. Extensive experiments demonstrate that VLMs trained with MedGround-35K consistently achieve improved referring grounding performance, enhance multi-object semantic disambiguation, and exhibit strong generalization to unseen grounding settings. This work highlights MedGround as a scalable, data-driven approach to anchor medical language to verifiable visual evidence.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
