StainFuser: Controlling Diffusion for Faster Neural Style Transfer in Multi-Gigapixel Histology Images

TL;DR

StainFuser introduces a novel diffusion-based style transfer method for rapid, high-quality stain normalization in multi-gigapixel histology images, significantly improving consistency and downstream analysis performance.

Contribution

The paper presents StainFuser, a new Conditional Latent Diffusion model for stain normalization, and introduces SPI-2M, the largest stain normalization dataset to date.

Findings

Outperforms existing methods in image quality

Enhances downstream model performance

Handles over 2 million histology images

Abstract

Stain normalization algorithms aim to transform the color and intensity characteristics of a source multi-gigapixel histology image to match those of a target image, mitigating inconsistencies in the appearance of stains used to highlight cellular components in the images. We propose a new approach, StainFuser, which treats this problem as a style transfer task using a novel Conditional Latent Diffusion architecture, eliminating the need for handcrafted color components. With this method, we curate SPI-2M the largest stain normalization dataset to date of over 2 million histology images with neural style transfer for high-quality transformations. Trained on this data, StainFuser outperforms current state-of-the-art deep learning and handcrafted methods in terms of the quality of normalized images and in terms of downstream model performance on the CoNIC dataset.