MetricUNet: Synergistic image- and voxel-level learning for precise prostate segmentation via online sampling |
| |
| Affiliation: | 1. Medical School of Nanjing University, Nanjing, China;2. National Institute of Healthcare Data Science at Nanjing University, Nanjing, China;3. School of Mathematics and Statistics, Xi’an Jiaotong University, Shanxi, China;4. Department of Psychiatry and Behavioral Sciences and the Department of Computer Science, Stanford University, CA, USA;5. State Key Laboratory for Novel Software Technology, Nanjing University, Nanjing, China;6. Department of Radiology, Nanjing Drum Tower Hospital, Nanjing University Medical School, Nanjing, China;7. School of Biomedical Engineering, ShanghaiTech University, Shanghai, China;8. Department of Research and Development, Shanghai United Imaging Intelligence Co., Ltd., Shanghai, China;9. Department of Artificial Intelligence, Korea University, Seoul 02841, Republic of Korea;1. College of Information Science and Technology, Dalian Maritime University, Dalian 116023, China;2. Department of Radiology and BRIC, University of North Carolina, Chapel Hill, NC 27599, USA;3. Departments of Psychiatry & Behavioral Sciences and Computer Science, Stanford University, Stanford, CA 94305-5723, USA;4. College of Marine Electrical Engineering, Dalian Maritime University, Dalian 116023, China;5. Department of Brain and Cognitive Engineering, Korea University, Seoul 02841, Republic of Korea;1. Beijing Engineering Research Center of Mixed Reality and Advanced Display, School of Optics and Photonics, Beijing Institute of Technology, Beijing 100081, China;2. Department of Radiology and BRIC, University of North Carolina at Chapel Hill, Chapel Hill, NC 27599, USA;3. School of Software, Beijing Institute of Technology, Beijing 100081, China;4. Department of Brain and Cognitive Engineering, Korea University, Seoul 02841, Republic of Korea |
| |
| Abstract: | Fully convolutional networks (FCNs), including UNet and VNet, are widely-used network architectures for semantic segmentation in recent studies. However, conventional FCN is typically trained by the cross-entropy or Dice loss, which only calculates the error between predictions and ground-truth labels for pixels individually. This often results in non-smooth neighborhoods in the predicted segmentation. This problem becomes more serious in CT prostate segmentation as CT images are usually of low tissue contrast. To address this problem, we propose a two-stage framework, with the first stage to quickly localize the prostate region, and the second stage to precisely segment the prostate by a multi-task UNet architecture. We introduce a novel online metric learning module through voxel-wise sampling in the multi-task network. Therefore, the proposed network has a dual-branch architecture that tackles two tasks: (1) a segmentation sub-network aiming to generate the prostate segmentation, and (2) a voxel-metric learning sub-network aiming to improve the quality of the learned feature space supervised by a metric loss. Specifically, the voxel-metric learning sub-network samples tuples (including triplets and pairs) in voxel-level through the intermediate feature maps. Unlike conventional deep metric learning methods that generate triplets or pairs in image-level before the training phase, our proposed voxel-wise tuples are sampled in an online manner and operated in an end-to-end fashion via multi-task learning. To evaluate the proposed method, we implement extensive experiments on a real CT image dataset consisting 339 patients. The ablation studies show that our method can effectively learn more representative voxel-level features compared with the conventional learning methods with cross-entropy or Dice loss. And the comparisons show that the proposed method outperforms the state-of-the-art methods by a reasonable margin. |
| |
| Keywords: | |
| 本文献已被 ScienceDirect 等数据库收录！ |
|
