Principled method cuts pathology image data size while boosting accuracy
Nonparametric Distribution Matching for Self-Supervised Whole-Slide Image Condensation
Machine Learning
Summary
Whole-slide images used in pathology are extremely large and hard for computers to handle. Prior methods tried to shrink these images for easier analysis but didn't keep features important for learning. The authors developed a new approach called NICER that reduces image data by matching feature distributions more precisely. Tests on multiple datasets and expert review showed NICER improved accuracy by about 7% and balanced speed and precision better than older methods.
What this means in practice
- •For medical imaging teams: Process large pathology slides faster with smaller data representations while keeping diagnostic features for machine learning tasks.
- •For digital pathology software developers: Integrate NICER’s condensation technique to improve whole-slide image analysis tools’ efficiency without losing critical information.
Authors
Duong M. Nguyen, Trong Nghia Hoang, Hang Thi Nguyen, Thanh Trung Huynh, Phi Le Nguyen, Minh N. Do
Abstract
Histological whole-slide images (WSIs) are central to computational pathology but pose severe computational challenges due to their extremely high resolution, often spanning several gigabytes per slide. To enable scalable learning, existing methods apply self-supervised data condensation to reduce computational cost, but typically rely on heuristic prototype learning and do not explicitly preserve learning-relevant feature distributions for downstream tasks. In response, we introduce a principled reformulation of WSI condensation as a distribution-matching problem under a fixed representational lens, and develop NICER, a tractable approximation framework based on a nonparametric prior with slide-adaptive capacity. Experiments on five histopathology datasets, together with clinical evaluation from a board-certified pathologist, show that NICER consistently outperforms prior methods, achieving an average accuracy improvement of 7.44% while offering improved efficiency-accuracy trade-offs, highlighting the benefits of principled, distribution-aware condensation for scalable histological representation learning. Source codes are available in https://github.com/nmduonggg/NICER.