Physics-Grounded Adversarial Stain Augmentation with Calibrated Coverage Guarantees

Mingi Hong

Keywords: Histopathology, Domain Generalization, Adversarial Augmentation, Distributionally Robust Optimization

TL;DR: Adversarial augmentation in the Macenko stain parameter space with a data-calibrated budget achieves 93.9%±1.4% on Camelyon17-WILDS—first in accuracy and lowest in variance among 10 methods.

Registration Requirement: Yes

Abstract:

Stain variation across hospitals degrades histopathology models at deployment.
Existing augmentation methods perturb color spaces with arbitrary hyperparameters, lacking both a principled budget and coverage guarantees for unseen centers.
We propose \textbf{C}alibrated \textbf{A}dversarial \textbf{S}tain \textbf{A}ugmentation (\textbf{CASA}), which performs adversarial augmentation in the Macenko stain parameter space with a budget calibrated from multi-center statistics via the DKW inequality.
On Camelyon17-WILDS (5 seeds), CASA achieves $93.9\% \pm 1.6\%$ slide-level accuracy---outperforming HED-strong ($88.4\% \pm 7.3\%$), RandStainNA ($85.2\% \pm 6.7\%$), and ERM ($63.9\% \pm 11.3\%$)---with the highest worst-group accuracy ($84.9\% \pm 0.9\%$) among all 10 compared methods.

Visa & Travel: No

Read CFP & Author Instructions: Yes

Originality Policy: Yes

Single-blind & Not Under Review Elsewhere: Yes

LLM Policy: Yes

Submission Number: 12