Margin Preserving Self-paced Contrastive Learning Towards Domain Adaptation for Medical Image Segmentation

TL;DR

This paper introduces a novel margin preserving self-paced contrastive learning approach for unsupervised domain adaptation in medical image segmentation, leveraging class-level prototypes and pseudo-labels to improve discriminability and domain-invariant features.

Contribution

The proposed MPSCL method innovatively combines margin preserving contrastive loss with self-paced pseudo-label generation for enhanced domain adaptation in medical imaging.

Findings

Significantly outperforms state-of-the-art methods on cross-modal cardiac segmentation.

Improves discriminability of feature representations through margin preserving contrastive loss.

Enhances domain-invariant feature learning with joint contrastive learning and pseudo-labels.

Abstract

To bridge the gap between the source and target domains in unsupervised domain adaptation (UDA), the most common strategy puts focus on matching the marginal distributions in the feature space through adversarial learning. However, such category-agnostic global alignment lacks of exploiting the class-level joint distributions, causing the aligned distribution less discriminative. To address this issue, we propose in this paper a novel margin preserving self-paced contrastive Learning (MPSCL) model for cross-modal medical image segmentation. Unlike the conventional construction of contrastive pairs in contrastive learning, the domain-adaptive category prototypes are utilized to constitute the positive and negative sample pairs. With the guidance of progressively refined semantic prototypes, a novel margin preserving contrastive loss is proposed to boost the discriminability of embedded representation space. To enhance the supervision for contrastive learning, more informative pseudo-labels are generated in target domain in a self-paced way, thus benefiting the category-aware distribution alignment for UDA. Furthermore, the domain-invariant representations are learned through joint contrastive learning between the two domains. Extensive experiments on cross-modal cardiac segmentation tasks demonstrate that MPSCL significantly improves semantic segmentation performance, and outperforms a wide variety of state-of-the-art methods by a large margin.