Robust Image Segmentation Quality Assessment
This addresses the need for reliable segmentation quality prediction in clinical applications, though it is incremental as it builds on existing regression-based methods to enhance robustness.
The paper tackles the problem of robust quality assessment for deep learning-based image segmentation without ground truth, particularly in clinical settings, by using the difference between input and reconstructed images to reduce overfitting and improve robustness, showing promising results on the ACDC17 dataset.
Deep learning based image segmentation methods have achieved great success, even having human-level accuracy in some applications. However, due to the black box nature of deep learning, the best method may fail in some situations. Thus predicting segmentation quality without ground truth would be very crucial especially in clinical practice. Recently, people proposed to train neural networks to estimate the quality score by regression. Although it can achieve promising prediction accuracy, the network suffers robustness problem, e.g. it is vulnerable to adversarial attacks. In this paper, we propose to alleviate this problem by utilizing the difference between the input image and the reconstructed image, which is conditioned on the segmentation to be assessed, to lower the chance to overfit to the undesired image features from the original input image, and thus to increase the robustness. Results on ACDC17 dataset demonstrated our method is promising.