We propose a novel semi-supervised image segmentation method that simultaneously optimizes a supervised segmentation and an unsupervised reconstruction objectives. The reconstruction objective uses an attention mechanism that separates the reconstruction of image areas corresponding to different classes. The proposed approach was evaluated on two applications: brain tumor and white matter hyperintensities segmentation. Our method, trained on unlabeled and a small number of labeled images, outperformed supervised CNNs trained with the same number of images and CNNs pre-trained on unlabeled data. In ablation experiments, we observed that the proposed attention mechanism substantially improves segmentation performance. We explore two multi-task training strategies: joint training and alternating training. Alternating training requires fewer hyperparameters and achieves a better, more stable performance than joint training. Finally, we analyze the features learned by different methods and find that the attention mechanism helps to learn more discriminative features in the deeper layers of encoders.

Additional Metadata
Keywords Attention, Brain tumor, Deep learning, Multi-task learning, Segmentation, Semi-supervised learning, White matter hyperintensities
Persistent URL dx.doi.org/10.1007/978-3-030-32248-9_51, hdl.handle.net/1765/122182
Series Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Citation
Chen, S. (Shuai), Bortsova, G, Juárez, A.G.-U. (Antonio García-Uceda), van Tulder, G, & de Bruijne, M. (2019). Multi-task Attention-Based Semi-supervised Learning for Medical Image Segmentation. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). doi:10.1007/978-3-030-32248-9_51