- [New] We further annotated an open available challenging cases dataset to evaluate the robustness and generalization of deep learning methods. Please check this repo RAOS.
- The real clinical application and assessment were conducted in this clinical paper, the code is available.
- This repo provides the codebase and dataset of work WORD: A large scale dataset, benchmark and clinical applicable study for abdominal organ segmentation from CT image. Each download requirement will be approved within two days.
- Now, we corrected the results of ESPNet+ KD in Table 8 and the dataset descriptions in Table 1 with red font Arxiv and LaTex.
- Some information about the WORD dataset is presented in the following (the LaTex style tables are here):
Fig. 1. An example in the WORD dataset.
Fig. 2. Volume distribution or each organ in the WORD dataset.
Fig. 3. Comparison results of CNN-based and Transformer-based methods.
Fig. 4. User study based on three junior oncologists independently, each of them comes from a different hospital.
DataSet
- Now, we have removed the download requirement permission, (PWD for BaiduPan is ABOD) and the WORD dataset unzip password is word@uestc, as we built a new robustness evaluation benchmark, please check this repo RAOS.
Please contact Xiangde (luoxd1996 AT gmail DOT com) for the dataset (the label of the testing set can be downloaded now labelTs). Two steps are needed to download and access the dataset: 1) using your google email to apply for the download permission (Goole Driven, BaiduPan); 2) using your affiliation email to get the unzip password/BaiduPan access code. We will get back to you within two days, so please don't send them multiple times. We just handle the real-name email and your email suffix must match your affiliation. The email should contain the following information:
~~Name/Homepage/Google Scholar: (Tell us who you are.)~~
~~Primary Affiliation: (The name of your institution or university, etc.)~~
~~Job Title: (E.g., Professor, Associate Professor, Ph.D., etc.)~~
~~Affiliation Email: (the password will be sent to this email, we just reply to the email which is the end of "edu".)~~
~~How to use: (Only for academic research, not for commercial use or second-development.)~~Acknowledgment and Statement
- This dataset belongs to the Healthcare Intelligence Laboratory at University of Electronic Science and Technology of China and is licensed under the GNU General Public License v3.0.
- This project has been approved by the privacy and ethical review committee. We thank all collaborators for the data collection, annotation, checking, and user study!
- This project and dataset were designed for open-available academic research, not for clinical, commercial, second-development, or other use. In addition, if you used it for your academic research, you are encouraged to release the code and the pre-trained model.
- The interesting and memorable name WORD is suggested by Dr. Jie-Neng, thanks a lot !!!
Citation
It would be highly appreciated if you cite our paper when using the WORD dataset or code:
@article{luo2022word,
title={{WORD}: A large scale dataset, benchmark and clinical applicable study for abdominal organ segmentation from CT image},
author={Xiangde Luo, Wenjun Liao, Jianghong Xiao, Jieneng Chen, Tao Song, Xiaofan Zhang, Kang Li, Dimitris N. Metaxas, Guotai Wang, and Shaoting Zhang},
journal={Medical Image Analysis},
volume={82},
pages={102642},
year={2022},
publisher={Elsevier}}
@article{liao2023comprehensive,
title={Comprehensive Evaluation of a Deep Learning Model for Automatic Organs-at-Risk Segmentation on Heterogeneous Computed Tomography Images for Abdominal Radiation Therapy},
author={Liao, Wenjun and Luo, Xiangde and He, Yuan and Dong, Ye and Li, Churong and Li, Kang and Zhang, Shichuan and Zhang, Shaoting and Wang, Guotai and Xiao, Jianghong},
journal={International Journal of Radiation Oncology* Biology* Physics},
volume={117},
number={4},
pages={994--1006},
year={2023},
publisher={Elsevier}}