Unsupervised Domain Adaptation Using Feature Disentanglement And GCNs For Medical Image Classification
This addresses generalization issues in medical image classification for healthcare applications, but it is incremental as it builds on existing domain adaptation techniques.
The paper tackles the problem of deep learning models failing to generalize due to distribution shifts in medical image data by proposing an unsupervised domain adaptation method using feature disentanglement and graph neural networks. It achieves state-of-the-art results on chest X-ray and histopathology datasets.
The success of deep learning has set new benchmarks for many medical image analysis tasks. However, deep models often fail to generalize in the presence of distribution shifts between training (source) data and test (target) data. One method commonly employed to counter distribution shifts is domain adaptation: using samples from the target domain to learn to account for shifted distributions. In this work we propose an unsupervised domain adaptation approach that uses graph neural networks and, disentangled semantic and domain invariant structural features, allowing for better performance across distribution shifts. We propose an extension to swapped autoencoders to obtain more discriminative features. We test the proposed method for classification on two challenging medical image datasets with distribution shifts - multi center chest Xray images and histopathology images. Experiments show our method achieves state-of-the-art results compared to other domain adaptation methods.