ASC-Net : Adversarial-based Selective Network for Unsupervised Anomaly Segmentation
This work addresses the problem of anomaly segmentation in medical imaging for applications with unlabelled data, representing an incremental advancement by bridging cluster-based and adversarial-based methods.
The paper tackles unsupervised anomaly segmentation by introducing ASC-Net, which uses adversarial learning to partition images based on user-defined reference distributions, achieving significant performance gains over methods like AnoGAN on tasks such as brain tumor and liver lesion segmentation.
We introduce a neural network framework, utilizing adversarial learning to partition an image into two cuts, with one cut falling into a reference distribution provided by the user. This concept tackles the task of unsupervised anomaly segmentation, which has attracted increasing attention in recent years due to their broad applications in tasks with unlabelled data. This Adversarial-based Selective Cutting network (ASC-Net) bridges the two domains of cluster-based deep learning methods and adversarial-based anomaly/novelty detection algorithms. We evaluate this unsupervised learning model on BraTS brain tumor segmentation, LiTS liver lesion segmentation, and MS-SEG2015 segmentation tasks. Compared to existing methods like the AnoGAN family, our model demonstrates tremendous performance gains in unsupervised anomaly segmentation tasks. Although there is still room to further improve performance compared to supervised learning algorithms, the promising experimental results shed light on building an unsupervised learning algorithm using user-defined knowledge.