Uncertainty Driven Bottleneck Attention U-Net For Organ at Risk Segmentation
Published in 2024 IEEE International Symposium on Biomedical Imaging (ISBI), 2024
This paper introduces a multiple decoder U-net architecture for organ-at-risk (OAR) segmentation in CT images, utilizing segmentation disagreement as attention and a CT intensity integrated regularization loss for enhanced accuracy. The proposed model shows improved performance on two OAR challenge datasets.
Recommended citation: A. Nazib, R. Hassan, Z. Islam and C. Fookes, "Uncertainty Driven Bottleneck Attention U-Net For Organ at Risk Segmentation," 2024 IEEE International Symposium on Biomedical Imaging (ISBI), Athens, Greece, 2024, pp. 1-5, doi: 10.1109/ISBI56570.2024.10635587. https://ieeexplore.ieee.org/document/10635587