English(EN) AGE-MIL: Anchor-Guided Evidence Learning for Patient-Level Prediction

新的AGE-MIL框架提升了病理学中的患者级别预测能力

作者 PulseAugur 编辑部 · [2 个来源] · 2026-06-10 14:19

研究人员推出了一种新颖的AGE-MIL框架，旨在改进计算病理学中的患者级别预测。这种弱监督方法解决了现有的全切片图像（WSI）级别方法与病理学家为诊断整合多张切片证据的方式之间的不匹配问题。AGE-MIL构建了一个患者级别的锚点，以捕捉全局上下文并指导相关局部块的整合，从而提高预测的可靠性。 AI

影响通过更好地整合多切片证据，提高了病理学中的诊断和预后准确性。

排序理由该集群包含一篇详细介绍计算病理学中患者级别预测新方法的学术论文。

AI 生成摘要 · Google Gemini · 来自 2 个来源。我们如何撰写摘要 →

报道来源 [2]

arXiv cs.CV TIER_1 English(EN) · Jiawei Niu, Jian Chen, Di Zhang, Junbo Lu, Zhangcheng Liao, Xuhao Liu, Honglin Zhong, Mireia Crispin-Ortuzar, Chen Li, Zeyu Gao, Yi Cai · 2026-06-11 04:00

AGE-MIL: Anchor-Guided Evidence Learning for Patient-Level Prediction

arXiv:2606.12126v1 Announce Type: new Abstract: Existing computational pathology methods predominantly operate within whole-slide image (WSI)-level multiple instance learning (MIL) paradigms, while patient-level modeling remains underexplored. In routine pathological practice, ho…
arXiv cs.CV TIER_1 English(EN) · Yi Cai · 2026-06-10 14:19

AGE-MIL: Anchor-Guided Evidence Learning for Patient-Level Prediction

Existing computational pathology methods predominantly operate within whole-slide image (WSI)-level multiple instance learning (MIL) paradigms, while patient-level modeling remains underexplored. In routine pathological practice, however, pathologists derive diagnostic and progno…