Improving Mitosis Detection Via UNet-based Adversarial Domain Homogenizer

TL;DR

This paper introduces a U-Net-based adversarial domain homogenizer to improve the robustness of mitosis detection in histology images, reducing domain bias and enhancing detection accuracy across different datasets.

Contribution

The paper presents a novel domain homogenizer that mitigates domain biases in histology images, improving the generalization of mitosis detection models.

Findings

Reduced domain differences in preprocessed images

Outperformed 2021 MIDOG challenge baselines in average precision

Effective adversarial reconstruction enhances detection robustness

Abstract

The effective localization of mitosis is a critical precursory task for deciding tumor prognosis and grade. Automated mitosis detection through deep learning-oriented image analysis often fails on unseen patient data due to inherent domain biases. This paper proposes a domain homogenizer for mitosis detection that attempts to alleviate domain differences in histology images via adversarial reconstruction of input images. The proposed homogenizer is based on a U-Net architecture and can effectively reduce domain differences commonly seen with histology imaging data. We demonstrate our domain homogenizer's effectiveness by observing the reduction in domain differences between the preprocessed images. Using this homogenizer, along with a subsequent retina-net object detector, we were able to outperform the baselines of the 2021 MIDOG challenge in terms of average precision of the detected mitotic figures.