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

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. INDEX TERMS Biomedical image segmentation, cross-modality learning, unsupervised domain adaptation, category-centric prototype aligner.

Design and Implementation of Convolutional Neural Networks Accelerator Based on Multidie

“…As one of the most representative networks in deep learning, convolutional neural networks (CNNs) achieved good performance in object detection [1], [2], image classification [3], [4], image segmentation [5], [6], and other fields. To capture more features, the structure of the model becomes larger, which leads to the high computational complexity of CNN.…”