# Weakly-supervised learning for medical image segmentation (WSL4MIS).
* This project was originally developed for our two previous works **[WORD](https://www.sciencedirect.com/science/article/pii/S1361841522002705)** (**MedIA2022**) and **[WSL4MIS](https://link.springer.com/chapter/10.1007/978-3-031-16431-6_50)** (**MICCAI2022**). If you use this project in your research, please cite the following works:
@article{luo2022scribbleseg,
title={Scribble-Supervised Medical Image Segmentation via Dual-Branch Network and Dynamically Mixed Pseudo Labels Supervision},
author={Xiangde Luo, Minhao Hu, Wenjun Liao, Shuwei Zhai, Tao Song, Guotai Wang, Shaoting Zhang},
journal={Medical Image Computing and Computer Assisted Intervention -- MICCAI 2022},
year={2022},
pages={528--538}}
@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}}
@misc{wsl4mis2020,
title={{WSL4MIS}},
author={Luo, Xiangde},
howpublished={\url{https://github.com/Luoxd1996/WSL4MIS}},
year={2021}}
* A re-implementation of this work based on the [PyMIC](https://github.com/HiLab-git/PyMIC) can be found here ([WSLDMPLS](https://github.com/HiLab-git/PyMIC_examples/tree/main/seg_wsl/ACDC)).
# Dataset
* The ACDC dataset with mask annotations can be downloaded from: [ACDC](https://www.creatis.insa-lyon.fr/Challenge/acdc/databases.html).
* The Scribble annotations of ACDC can be downloaded from: [Scribble](https://gvalvano.github.io/wss-multiscale-adversarial-attention-gates/data).
* The data processing code in [Here](https://github.com/Luoxd1996/WSL4MIS/blob/main/code/dataloaders/acdc_data_processing.py) the pre-processed ACDC data in [Here](https://github.com/HiLab-git/WSL4MIS/tree/main/data/ACDC).
* The ISBI-MR-Prostate-2013 dataset with mask annotation can be downloaded from [TCIA](https://wiki.cancerimagingarchive.net/pages/viewpage.action?pageId=21267207), the scribble annotation of this dataset (annotated by Mr. X. Luo and M.D. W. Liao) can be downloaded [GoogleDrive](https://drive.google.com/file/d/1VFKP1-bychADhGw5rbRtd_JKLallccGz/view?usp=sharing) and [BaiduPan](https://pan.baidu.com/s/1jLz5tDAUBw4deKxHntv0gg?pwd=jqr1).
* **To simulate the scribble annotation for other datasets, we further provide the simulation code at [Here](https://github.com/HiLab-git/WSL4MIS/blob/main/code/scribbles_generator.py)**.
# Requirements
Some important required packages include:
* [Pytorch][torch_link] version >=0.4.1.
* TensorBoardX
* Python >= 3.6
* Efficientnet-Pytorch `pip install efficientnet_pytorch`
* Some basic python packages such as Numpy, Scikit-image, SimpleITK, Scipy ......
Follow official guidance to install [Pytorch][torch_link].
[torch_link]:https://pytorch.org/
# Usage
1. Clone this project.
```
git clone https://github.com/HiLab-git/WSL4MIS
cd WSL4MIS
```
2. Data pre-processing os used or the processed data.
```
cd code
python dataloaders/acdc_data_processing.py
```
3. Train the model
```
cd code
bash train_wss.sh # train model with scribble or dense annotations.
bash train_ssl.sh # train model with mix-supervision (mask annotations and without annotation).
```
4. Test the model
```
python test_2D_fully.py --sup_type scribble/label --exp ACDC/the trained model fold --model unet
python test_2D_fully_sps.py --sup_type scribble --exp ACDC/the trained model fold --model unet_cct
```
5. Training curves on the fold1:
![](https://github.com/Luoxd1996/WSL4MIS/blob/main/imgs/fold1_curve.png)
**Note**: pCE means partially cross-entropy, TV means total variation, label denotes supervised by mask, scribble represents just supervised by scribbles.
# Implemented methods
* [**pCE**](https://openaccess.thecvf.com/content_cvpr_2018/papers/Tang_Normalized_Cut_Loss_CVPR_2018_paper.pdf)
* [**pCE + TV**](https://arxiv.org/pdf/1605.01368.pdf)
* [**pCE + Entropy Minimization**](https://arxiv.org/pdf/2111.02403.pdf)
* [**pCE + GatedCRFLoss**](https://github.com/LEONOB2014/GatedCRFLoss)
* [**pCE + Intensity Variance Minimization**](https://arxiv.org/pdf/2111.02403.pdf)
* [**pCE + Random Walker**](http://vision.cse.psu.edu/people/chenpingY/paper/grady2006random.pdf)
* [**pCE + MumfordShah_Loss**](https://arxiv.org/pdf/1904.02872.pdf)
* [**Scribble2Label**](https://arxiv.org/pdf/2006.12890.pdf)
* [**USTM**](https://www.sciencedirect.com/science/article/pii/S0031320321005215)
* [**ScribbleVC**](https://github.com/HUANGLIZI/ScribbleVC)
# Acknowledgement
* The GatedCRFLoss is adapted from [GatedCRFLoss](https://github.com/LEONOB2014/GatedCRFLoss) for medical image segmentation.
* The codebase is adapted from our previous work [SSL4MIS](https://github.com/HiLab-git/SSL4MIS).
* The WORD dataset will be presented at [WORD](https://github.com/HiLab-git/WORD).
没有合适的资源?快使用搜索试试~ 我知道了~
温馨提示
本项目是demo,可以直接使用代码注释清楚,文档教程完整。 医学图像分割是一项重要的计算机视觉任务,它涉及到从医学图像中准确地分割出感兴趣的区域,如器官、病变等。弱监督学习在医学图像分割中是一个新兴的研究领域,因为它允许使用有限的标注数据来训练模型,这对于医疗图像分析来说是一个巨大的优势,因为在实际应用中获取大量高质量的标注数据是非常困难的和昂贵的。
资源推荐
资源详情
资源评论
收起资源包目录
医学图像分割的弱监督学习.zip (2000个子文件)
patient057_frame09.h5 1.51MB
patient057_frame01.h5 1.51MB
patient006_frame01.h5 1MB
patient006_frame16.h5 1022KB
patient038_frame11.h5 1008KB
patient038_frame01.h5 1008KB
patient004_frame15.h5 967KB
patient004_frame01.h5 962KB
patient035_frame11.h5 914KB
patient035_frame01.h5 909KB
patient085_frame09.h5 908KB
patient085_frame01.h5 881KB
patient008_frame01.h5 842KB
patient008_frame13.h5 840KB
patient081_frame01.h5 783KB
patient025_frame01.h5 774KB
patient099_frame01.h5 773KB
patient099_frame09.h5 773KB
patient025_frame09.h5 772KB
patient082_frame01.h5 768KB
patient082_frame07.h5 764KB
patient081_frame07.h5 763KB
patient050_frame01.h5 757KB
patient050_frame12.h5 755KB
patient002_frame12.h5 754KB
patient088_frame01.h5 753KB
patient088_frame12.h5 752KB
patient002_frame01.h5 752KB
patient032_frame01.h5 745KB
patient032_frame12.h5 741KB
patient003_frame15.h5 732KB
patient075_frame01.h5 731KB
patient003_frame01.h5 730KB
patient075_frame06.h5 729KB
patient042_frame16.h5 719KB
patient021_frame01.h5 718KB
patient021_frame13.h5 714KB
patient066_frame11.h5 709KB
patient066_frame01.h5 709KB
patient033_frame01.h5 707KB
patient042_frame01.h5 707KB
patient024_frame09.h5 704KB
patient024_frame01.h5 703KB
patient033_frame14.h5 703KB
patient064_frame12.h5 702KB
patient040_frame13.h5 696KB
patient005_frame01.h5 690KB
patient001_frame01.h5 689KB
patient001_frame12.h5 688KB
patient005_frame13.h5 686KB
patient100_frame13.h5 680KB
patient027_frame01.h5 680KB
patient040_frame01.h5 680KB
patient100_frame01.h5 679KB
patient027_frame11.h5 677KB
patient064_frame01.h5 677KB
patient031_frame10.h5 676KB
patient030_frame01.h5 676KB
patient030_frame12.h5 675KB
patient007_frame01.h5 675KB
patient023_frame01.h5 673KB
patient023_frame09.h5 665KB
patient084_frame10.h5 661KB
patient084_frame01.h5 660KB
patient031_frame01.h5 658KB
patient010_frame01.h5 654KB
patient063_frame01.h5 653KB
patient007_frame07.h5 653KB
patient073_frame01.h5 653KB
patient073_frame10.h5 651KB
patient063_frame16.h5 650KB
patient019_frame01.h5 647KB
patient095_frame12.h5 647KB
patient019_frame11.h5 646KB
patient096_frame08.h5 646KB
patient096_frame01.h5 645KB
patient095_frame01.h5 643KB
patient009_frame01.h5 634KB
patient026_frame01.h5 632KB
patient026_frame12.h5 631KB
patient009_frame13.h5 630KB
patient039_frame10.h5 630KB
patient039_frame01.h5 630KB
patient010_frame13.h5 629KB
patient061_frame01.h5 625KB
patient061_frame10.h5 624KB
patient043_frame07.h5 615KB
patient029_frame12.h5 614KB
patient029_frame01.h5 613KB
patient043_frame01.h5 609KB
patient092_frame01.h5 600KB
patient092_frame06.h5 599KB
patient034_frame16.h5 594KB
patient051_frame11.h5 594KB
patient034_frame01.h5 591KB
patient028_frame01.h5 590KB
patient051_frame01.h5 589KB
patient028_frame09.h5 585KB
patient067_frame10.h5 583KB
patient047_frame01.h5 583KB
共 2000 条
- 1
- 2
- 3
- 4
- 5
- 6
- 20
资源评论
小码蚁.
- 粉丝: 2534
- 资源: 4176
上传资源 快速赚钱
- 我的内容管理 展开
- 我的资源 快来上传第一个资源
- 我的收益 登录查看自己的收益
- 我的积分 登录查看自己的积分
- 我的C币 登录后查看C币余额
- 我的收藏
- 我的下载
- 下载帮助
安全验证
文档复制为VIP权益,开通VIP直接复制
信息提交成功