Unsupervised Histopathology Image Synthesis

TL;DR

This paper presents an unsupervised pipeline for synthesizing realistic and representative histopathology images, improving CNN training for nucleus segmentation across multiple cancer types without extensive labeled data.

Contribution

The novel unsupervised synthesis method generates realistic, style-referenced, and challenging images that enhance CNN training, outperforming supervised methods in data-scarce scenarios.

Findings

Unsupervised synthetic images improve nucleus segmentation performance.

Method outperforms supervised approaches when no labeled data is available.

Synthetic images are realistic, style-referenced, and challenging for training.

Abstract

Hematoxylin and Eosin stained histopathology image analysis is essential for the diagnosis and study of complicated diseases such as cancer. Existing state-of-the-art approaches demand extensive amount of supervised training data from trained pathologists. In this work we synthesize in an unsupervised manner, large histopathology image datasets, suitable for supervised training tasks. We propose a unified pipeline that: a) generates a set of initial synthetic histopathology images with paired information about the nuclei such as segmentation masks; b) refines the initial synthetic images through a Generative Adversarial Network (GAN) to reference styles; c) trains a task-specific CNN and boosts the performance of the task-specific CNN with on-the-fly generated adversarial examples. Our main contribution is that the synthetic images are not only realistic, but also representative (in reference styles) and relatively challenging for training task-specific CNNs. We test our method for nucleus segmentation using images from four cancer types. When no supervised data exists for a cancer type, our method without supervision cost significantly outperforms supervised methods which perform across-cancer generalization. Even when supervised data exists for all cancer types, our approach without supervision cost performs better than supervised methods.