Unsupervised Bidirectional Cross-Modality Adaptation via Deeply Synergistic Image and Feature Alignment for Medical Image Segmentation
This work addresses the problem of domain shift in medical imaging for clinicians and researchers, enabling more reliable segmentation across modalities like MRI and CT, but it is incremental as it builds on existing adversarial learning methods.
The authors tackled performance degradation in medical image segmentation when adapting deep networks to unlabeled target domains with heterogeneous characteristics, by proposing a novel unsupervised domain adaptation framework called SIFA that synergistically aligns domains from image and feature perspectives, resulting in improved segmentation performance and outperforming state-of-the-art approaches by a large margin.
Unsupervised domain adaptation has increasingly gained interest in medical image computing, aiming to tackle the performance degradation of deep neural networks when being deployed to unseen data with heterogeneous characteristics. In this work, we present a novel unsupervised domain adaptation framework, named as Synergistic Image and Feature Alignment (SIFA), to effectively adapt a segmentation network to an unlabeled target domain. Our proposed SIFA conducts synergistic alignment of domains from both image and feature perspectives. In particular, we simultaneously transform the appearance of images across domains and enhance domain-invariance of the extracted features by leveraging adversarial learning in multiple aspects and with a deeply supervised mechanism. The feature encoder is shared between both adaptive perspectives to leverage their mutual benefits via end-to-end learning. We have extensively evaluated our method with cardiac substructure segmentation and abdominal multi-organ segmentation for bidirectional cross-modality adaptation between MRI and CT images. Experimental results on two different tasks demonstrate that our SIFA method is effective in improving segmentation performance on unlabeled target images, and outperforms the state-of-the-art domain adaptation approaches by a large margin.