StainGAN: Stain Style Transfer for Digital Histological Images

TL;DR

StainGAN is a deep learning model that performs stain style transfer on histological images, improving color consistency without needing reference slides, and enhances clinical diagnosis accuracy.

Contribution

Introduces StainGAN, a CycleGAN-inspired deep learning approach for stain style transfer that eliminates the dependency on reference slides in histological image processing.

Findings

10% improvement in SSIM over state-of-the-art methods

12% increase in AUC for breast cancer classification

Validated on clinical breast cancer diagnosis

Abstract

Digitized Histological diagnosis is in increasing demand. However, color variations due to various factors are imposing obstacles to the diagnosis process. The problem of stain color variations is a well-defined problem with many proposed solutions. Most of these solutions are highly dependent on a reference template slide. We propose a deep-learning solution inspired by CycleGANs that is trained end-to-end, eliminating the need for an expert to pick a representative reference slide. Our approach showed superior results quantitatively and qualitatively against the state of the art methods (10% improvement visually using SSIM). We further validated our method on a clinical use-case, namely Breast Cancer tumor classification, showing 12% increase in AUC. The code will be made publicly available.