Comparing Adversarial and Supervised Learning for Organs at Risk Segmentation in CT images
This work addresses segmentation accuracy for radiotherapy planning, but it is incremental as it compares existing deep learning methods on a specific medical dataset.
The paper tackled the problem of automating Organ at Risk segmentation in CT images for radiotherapy by comparing GAN-based models with supervised CNN approaches, finding that the GAN models performed similarly or better, especially for challenging organs.
Organ at Risk (OAR) segmentation from CT scans is a key component of the radiotherapy treatment workflow. In recent years, deep learning techniques have shown remarkable potential in automating this process. In this paper, we investigate the performance of Generative Adversarial Networks (GANs) compared to supervised learning approaches for segmenting OARs from CT images. We propose three GAN-based models with identical generator architectures but different discriminator networks. These models are compared with well-established CNN models, such as SE-ResUnet and DeepLabV3, using the StructSeg dataset, which consists of 50 annotated CT scans containing contours of six OARs. Our work aims to provide insight into the advantages and disadvantages of adversarial training in the context of OAR segmentation. The results are very promising and show that the proposed GAN-based approaches are similar or superior to their CNN-based counterparts, particularly when segmenting more challenging target organs.