An all-in-one foundation model for multimodal medical image restoration and enhancement
摘要
Nature Machine Intelligence 在线发表 Cao 等人的研究，提出基础模型 AMIFound，将七种医学影像模态的修复任务统一到一个模型中，可在多种退化条件下完成增强，并保留对诊断有临床意义的信息。论文为预印本之外的期刊在线发表版本，DOI 为 10.1038/s42256-026-01310-z。
摘要由模型根据原文撰写，以原文为准。
正文 · 原文
Nature Machine Intelligence, Published online: 13 October 2026; doi:10.1038/s42256-026-01310-z
Cao et al. present AMIFound, a foundation model that unifies medical image restoration across seven imaging modalities, delivering robust enhancement across diverse degradations and preserving clinically relevant information for diagnosis.
注意转载授权 · 阅读原文
来源：Nature Machine Intelligence · nature.com
引用原文
Luyang Cao. An all-in-one foundation model for multimodal medical image restoration and enhancement[EB/OL]. (2026-10-13)[访问日期]. https://www.nature.com/articles/s42256-026-01310-z.