Collaborative Unsupervised Domain Adaptation for Medical Image Diagnosis

TL;DR

This paper introduces a collaborative unsupervised domain adaptation method for medical image diagnosis that effectively handles limited and noisy labels, improving model transferability and generalization across domains.

Contribution

It proposes a novel transferability-aware adaptation algorithm that addresses label noise and domain differences in medical image diagnosis tasks.

Findings

The method outperforms existing UDA techniques on medical image datasets.

The approach demonstrates robustness to label noise in training data.

Theoretical analysis confirms improved generalization performance.

Abstract

Deep learning based medical image diagnosis has shown great potential in clinical medicine. However, it often suffers two major difficulties in real-world applications: 1) only limited labels are available for model training, due to expensive annotation costs over medical images; 2) labeled images may contain considerable label noise (e.g., mislabeling labels) due to diagnostic difficulties of diseases. To address these, we seek to exploit rich labeled data from relevant domains to help the learning in the target task via {Unsupervised Domain Adaptation} (UDA). Unlike most UDA methods that rely on clean labeled data or assume samples are equally transferable, we innovatively propose a Collaborative Unsupervised Domain Adaptation algorithm, which conducts transferability-aware adaptation and conquers label noise in a collaborative way. We theoretically analyze the generalization performance of the proposed method, and also empirically evaluate it on both medical and general images. Promising experimental results demonstrate the superiority and generalization of the proposed method.