Stain Isolation-based Guidance for Improved Stain Translation

TL;DR

This paper introduces a stain isolation guidance scheme for stain translation in histopathology images, improving structure preservation in unsupervised domain translation methods like CycleGAN.

Contribution

It proposes a novel guidance loss based on stain isolation that enhances structure preservation during stain translation.

Findings

Improved translation quality demonstrated qualitatively.

Quantitative results show better structure preservation.

Effective for IHC and mIF stain translation.

Abstract

Unsupervised and unpaired domain translation using generative adversarial neural networks, and more precisely CycleGAN, is state of the art for the stain translation of histopathology images. It often, however, suffers from the presence of cycle-consistent but non structure-preserving errors. We propose an alternative approach to the set of methods which, relying on segmentation consistency, enable the preservation of pathology structures. Focusing on immunohistochemistry (IHC) and multiplexed immunofluorescence (mIF), we introduce a simple yet effective guidance scheme as a loss function that leverages the consistency of stain translation with stain isolation. Qualitative and quantitative experiments show the ability of the proposed approach to improve translation between the two domains.