Domain and Content Adaptive Convolutions for Cross-Domain Adenocarcinoma Segmentation

TL;DR

This paper introduces a U-Net-based segmentation framework with domain and content adaptive convolutions, effectively addressing cross-domain adenocarcinoma segmentation challenges caused by morphological and scanner variations, achieving top challenge scores.

Contribution

The paper proposes a novel adaptive convolutional approach tailored for cross-domain histopathology segmentation, improving robustness against domain shifts.

Findings

Achieved segmentation scores of 0.8020 and 0.8527 on challenge tracks.

Ranked as the best-performing submission in the COSAS challenge.

Demonstrated robustness to morphological and scanner-induced domain shifts.

Abstract

Recent advances in computer-aided diagnosis for histopathology have been largely driven by the use of deep learning models for automated image analysis. While these networks can perform on par with medical experts, their performance can be impeded by out-of-distribution data. The Cross-Organ and Cross-Scanner Adenocarcinoma Segmentation (COSAS) challenge aimed to address the task of cross-domain adenocarcinoma segmentation in the presence of morphological and scanner-induced domain shifts. In this paper, we present a U-Net-based segmentation framework designed to tackle this challenge. Our approach achieved segmentation scores of 0.8020 for the cross-organ track and 0.8527 for the cross-scanner track on the final challenge test sets, ranking it the best-performing submission.