English(EN) Rethinking Noise-Robust Training for Frozen Vision Foundation Models: A Cross-Dataset Benchmark with a Case Study of Small-Loss Failure

视觉模型噪声鲁棒性基准显示无单一赢家

作者 PulseAugur 编辑部 · [2 个来源] · 2026-05-21 15:02

一项关于冻结视觉基础模型噪声鲁棒性训练的新基准研究表明，在各种医学影像数据集和噪声条件下，没有一种方法能够始终优于其他方法。研究强调，方法的选择对性能有显著影响，尤其是在噪声严重性增加的情况下。研究结果表明，根据特定的噪声模式选择合适的方法比寻找普遍占优的算法更为关键。 AI

影响强调了为视觉模型选择噪声鲁棒性训练方法的复杂性，表明需要根据特定模式进行选择，而不是依赖单一最佳算法。

排序理由学术论文，提出了一个新的基准和分析。

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) · Zitong Li, Haoyu Wang · 2026-05-22 04:00

Rethinking Noise-Robust Training for Frozen Vision Foundation Models: A Cross-Dataset Benchmark with a Case Study of Small-Loss Failure

arXiv:2605.22591v1 Announce Type: new Abstract: Frozen Vision Foundation Models (VFMs) with lightweight classification heads are increasingly used in medical imaging because they offer efficient and reproducible deployment. Yet noisy-label learning methods for this frozen-feature…
arXiv cs.CV TIER_1 English(EN) · Haoyu Wang · 2026-05-21 15:02

Rethinking Noise-Robust Training for Frozen Vision Foundation Models: A Cross-Dataset Benchmark with a Case Study of Small-Loss Failure

Frozen Vision Foundation Models (VFMs) with lightweight classification heads are increasingly used in medical imaging because they offer efficient and reproducible deployment. Yet noisy-label learning methods for this frozen-feature regime remain poorly understood, and most exist…