Consistent Posterior Distributions under Vessel-Mixing: A Regularization for Cross-Domain Retinal Artery/Vein Classification

TL;DR

This paper introduces a vessel-mixing regularization method to improve cross-domain retinal artery/vein classification, enhancing model robustness against domain shifts caused by different imaging protocols.

Contribution

The paper proposes a novel vessel-mixing based consistency regularization framework that improves cross-domain retinal A/V classification by enhancing model robustness to domain shifts.

Findings

Achieves state-of-the-art cross-domain performance.

Close to fully supervised upper bound on target domain.

Effective in handling diverse datasets from different institutions.

Abstract

Retinal artery/vein (A/V) classification is a critical technique for diagnosing diabetes and cardiovascular diseases. Although deep learning based methods achieve impressive results in A/V classification, their performances usually degrade severely when being directly applied to another database, due to the domain shift, e.g., caused by the variations in imaging protocols. In this paper, we propose a novel vessel-mixing based consistency regularization framework, for cross-domain learning in retinal A/V classification. Specially, to alleviate the severe bias to source domain, based on the label smooth prior, the model is regularized to give consistent predictions for unlabeled target-domain inputs that are under perturbation. This consistency regularization implicitly introduces a mechanism where the model and the perturbation is opponent to each other, where the model is pushed to be robust enough to cope with the perturbation. Thus, we investigate a more difficult opponent to further inspire the robustness of model, in the scenario of retinal A/V, called vessel-mixing perturbation. Specially, it effectively disturbs the fundus images especially the vessel structures by mixing two images regionally. We conduct extensive experiments on cross-domain A/V classification using four public datasets, which are collected by diverse institutions and imaging devices. The results demonstrate that our method achieves the state-of-the-art cross-domain performance, which is also close to the upper bound obtained by fully supervised learning on target domain.