One-Slide Calibration of Pathology Foundation Models

Key points

• We introduce SlideRuler, which uses regions within a slide as internal controls to estimate and correct acquisition-induced shifts in other regions.1
• A transfer map learned from paired rescans enables calibration from a single scan at inference while keeping the foundation model fixed.1
• 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.1
• By drawing calibration information from the slide itself, SlideRuler offers a path toward more consistent use of frozen pathology models across imaging systems.1

Sources (1)

• [1]One-Slide Calibration of Pathology Foundation Models
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 6, 06:09 PM

  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.

Extractive summary: sentences quoted from the sources.