CXR Segmentation by AdaIN-based Domain Adaptation and Knowledge Distillation
This work addresses domain shift in medical imaging segmentation, particularly for abnormal chest X-rays, but is incremental as it builds on existing multi-domain translation and knowledge distillation techniques.
The paper tackles the problem of chest X-ray segmentation with scarce labels by proposing a framework that combines domain adaptation and semi-supervised learning using AdaIN and knowledge distillation, achieving state-of-the-art performance for abnormal CXR segmentation.
As segmentation labels are scarce, extensive researches have been conducted to train segmentation networks with domain adaptation, semi-supervised or self-supervised learning techniques to utilize abundant unlabeled dataset. However, these approaches appear different from each other, so it is not clear how these approaches can be combined for better performance. Inspired by recent multi-domain image translation approaches, here we propose a novel segmentation framework using adaptive instance normalization (AdaIN), so that a single generator is trained to perform both domain adaptation and semi-supervised segmentation tasks via knowledge distillation by simply changing task-specific AdaIN codes. Specifically, our framework is designed to deal with difficult situations in chest X-ray radiograph (CXR) segmentation, where labels are only available for normal data, but the trained model should be applied to both normal and abnormal data. The proposed network demonstrates great generalizability under domain shift and achieves the state-of-the-art performance for abnormal CXR segmentation.