arXiv 2602.095872026-05-12cs.CVcs.AI

MieDB-100k：用于医学图像编辑的综合数据集

MieDB-100k: A Comprehensive Dataset for Medical Image Editing

State Key Laboratory of General Artificial Intelligence, Beijing, China（1 国家一般人工智能重点实验室，北京，中国）；School of Intelligence Science and Technology, Peking University, Beijing, China（2 智能科学与技术学院，北京大学，北京，中国）；National Institute of Health Data Science, Peking University, Beijing, China（3 国家健康数据科学研究院，北京大学，北京，中国）；DAMO Academy, Alibaba Group, Zhejiang, China（4 阿里巴巴集团 DAMO 院，浙江，中国）；hupan lab, zhejiang province（5 鹏元实验室，浙江省）；Zhejiang University, Zhejiang, China（6 浙江大学，浙江，中国）

Yongfan Lai, Wen Qian, Bo Liu, Hongyan Li, Hao Luo, Fan Wang, Bohan Zhuang, Shenda Hong

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

本文提出MieDB-100k数据集，通过数据筛选流程结合专家模型与规则生成方法，解决医学图像编辑中数据质量、多样性与可扩展性不足的问题，实验表明其在医学图像编辑任务中表现优异。

摘要：

英文摘要：

The scarcity of high-quality data remains a primary bottleneck in adapting multimodal generative models for medical image editing. Existing medical image editing datasets often suffer from limited diversity, neglect of medical image understanding and inability to balance quality with scalability. To address these gaps, we propose MieDB-100k, a large-scale, high-quality and diverse dataset for text-guided medical image editing. It categorizes editing tasks into perspectives of Perception, Modification and Transformation, considering both understanding and generation abilities. We construct MieDB-100k via a data curation pipeline leveraging both modality-specific expert models and rule-based data synthetic methods, followed by rigorous manual inspection to ensure clinical fidelity. Extensive experiments demonstrate that model trained with MieDB-100k consistently outperform both open-source and proprietary models while exhibiting strong generalization ability. We anticipate that this dataset will serve as a cornerstone for future advancements in specialized medical image editing.

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。