analiza aiqbench
This research addresses the critical challenge of domain shift in medical AI by introducing SlideRuler, a method that calibrates frozen pathology foundation models using internal slide controls to mitigate scanner-induced variations. By enabling consistent model performance across different imaging systems without retraining, this approach significantly advances the practical deployment of robust computer vision tools in clinical settings.
One-Slide Calibration of Pathology Foundation Models
arXiv:2610.08944v1 Announce Type: cross Abstract: Scanner variation changes how pathology foundation models represent the same tissue. We introduce SlideRuler, which uses regions within a slide as internal controls to estimate and correct acquisition-induced shifts in other regions. A transfer map learned from paired rescans enables calibration from a single scan at inference while keeping the foundation model fixed. Across two encoders and five SCORPION scanners, learned transfer reduces mean target-to-source embedding distance by 16.3-38.5% relative to raw embeddings. Comparisons with unrelated same-scanner controls reveal a positive same-slide contribution across all four evaluation settings, including scanner holdout. A source-anchored variant reduces source-feature displacement by 47.7-83.6% relative to learned transfer while retaining most of its alignment gain. By drawing calibration information from the slide itself, SlideRuler offers a path toward more consistent use of frozen pathology models across imaging systems.
Abstract:Scanner variation changes how pathology foundation models represent the same tissue. We introduce SlideRuler, which uses regions within a slide as internal controls to estimate and correct acquisition-induced shifts in other regions. A transfer map learned from paired rescans enables calibration from a single scan at inference while keeping the foundation model fixed. Across two encoders and five SCORPION scanners, learned transfer reduces mean target-to-source embedding distance by 16.3-38.5% relative to raw embeddings. Comparisons with unrelated same-scanner controls reveal a positive same-slide contribution across all four evaluation settings, including scanner holdout. A source-anchored variant reduces source-feature displacement by 47.7-83.6% relative to learned transfer while retaining most of its alignment gain. By drawing calibration information from the slide itself, SlideRuler offers a path toward more consistent use of frozen pathology models across imaging systems.
Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
arXiv:2610.08944 [cs.CV]
(or arXiv:2610.08944v1 [cs.CV] for this version)
https://doi.org/10.48550/arXiv.2610.08944 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 233 cuvinte
Subiecte conexe
Introducing NV-Reason-CT Open 3D CT VLM for Radiologist Chain-of-Thought Reasoning
NVIDIA · 23 sept. 2026
Presentation: Building Evals for AI Adoption: from Principles to Practice
InfoQ · 29 mai 2026
GigaPath-Flash and GigaTIME-Flash: Toward population-scale discovery with efficient pathology foundation models
Microsoft Research · 31 aug. 2026
GigaPath-Flash and GigaTIME-Flash: Efficient Pathology Foundation Models for Whole-Slide and Tumor Microenvironment Analysis
arXiv · 21 iul. 2026