Unsupervised Domain Adaptation Network with Category-Centric Prototype Aligner for Biomedical Image Segmentation

2021/03/03 by Ping Gong, Wenwen Yu, Gong, Ping +7 · 1 citation
Computer Science · Medicine · #Artificial Intelligence in Healthcare and Education #Artificial intelligence #COVID-19 diagnosis using AI #Classifier (UML) #Computer Vision and Pattern Recognition (cs.CV) #Computer science #Detector #Discriminator #Domain Adaptation and Few-Shot Learning #Exploit #FOS: Computer and information sciences #Image segmentation #Machine learning #Pattern recognition (psychology) #Segmentation #cs.CV

paper · pdf · doi:10.48550/arxiv.2103.02220

published in arXiv (Cornell University) (Cornell University) · Ping Gong and Wenwen Yu contributed equally to this work. 11 pages, 4 figures, 3 tables

arxiv created 2021/03/03 · openalex publication_date 2021/03/03 · arxiv updated 2021/03/04 · openalex created_date 2021/03/15 · openalex updated_date 2026/07/28

Abstract

With the widespread success of deep learning in biomedical image segmentation, domain shift becomes a critical and challenging problem, as the gap between two domains can severely affect model performance when deployed to unseen data with heterogeneous features. To alleviate this problem, we present a novel unsupervised domain adaptation network, for generalizing models learned from the labeled source domain to the unlabeled target domain for cross-modality biomedical image segmentation. Specifically, our approach consists of two key modules, a conditional domain discriminator~(CDD) and a category-centric prototype aligner~(CCPA). The CDD, extended from conditional domain adversarial networks in classifier tasks, is effective and robust in handling complex cross-modality biomedical images. The CCPA, improved from the graph-induced prototype alignment mechanism in cross-domain object detection, can exploit precise instance-level features through an elaborate prototype representation. In addition, it can address the negative effect of class imbalance via entropy-based loss. Extensive experiments on a public benchmark for the cardiac substructure segmentation task demonstrate that our method significantly improves performance on the target domain.

Citations

• Deep Domain Confusion: Maximizing for Domain Invariance
• Visual Domain Adaptation: A survey of recent advances
• Learning Transferable Features with Deep Adaptation Networks
• Dilated Residual Networks
• Conditional Adversarial Domain Adaptation
• ADVENT: Adversarial Entropy Minimization for Domain Adaptation in Semantic Segmentation
• Wasserstein Distance Guided Representation Learning for Domain Adaptation
• Unsupervised Domain Adaptation in Semantic Segmentation: a Review

Related

• Deep Symmetric Adaptation Network for Cross-modality Medical Image Segmentation
• Unsupervised Bidirectional Cross-Modality Adaptation via Deeply Synergistic Image and Feature Alignment for Medical Image Segmentation
• Unsupervised Cross-Modality Domain Adaptation of ConvNets for Biomedical Image Segmentations with Adversarial Loss
• Synergistic Image and Feature Adaptation: Towards Cross-Modality Domain Adaptation for Medical Image Segmentation
• PnP-AdaNet: Plug-and-Play Adversarial Domain Adaptation Network with a Benchmark at Cross-modality Cardiac Segmentation