A Laplacian Pyramid Based Generative H&E Stain Augmentation Network
This work addresses a domain-specific problem for medical diagnostics by providing an incremental improvement in stain augmentation for histology images.
The paper tackles the problem of stain variability in H&E histology images, which hinders generalization of machine-learning diagnostic tools, by proposing a GAN-based framework for stain augmentation that improves patch classification F1 score by 15.7% and nucleus segmentation panoptic quality by 7.3%.
Hematoxylin and Eosin (H&E) staining is a widely used sample preparation procedure for enhancing the saturation of tissue sections and the contrast between nuclei and cytoplasm in histology images for medical diagnostics. However, various factors, such as the differences in the reagents used, result in high variability in the colors of the stains actually recorded. This variability poses a challenge in achieving generalization for machine-learning based computer-aided diagnostic tools. To desensitize the learned models to stain variations, we propose the Generative Stain Augmentation Network (G-SAN) -- a GAN-based framework that augments a collection of cell images with simulated yet realistic stain variations. At its core, G-SAN uses a novel and highly computationally efficient Laplacian Pyramid (LP) based generator architecture, that is capable of disentangling stain from cell morphology. Through the task of patch classification and nucleus segmentation, we show that using G-SAN-augmented training data provides on average 15.7% improvement in F1 score and 7.3% improvement in panoptic quality, respectively. Our code is available at https://github.com/lifangda01/GSAN-Demo.