# TransBTS(MICCAI2021)& TransBTSV2 (To Be Updated)
This repo is the official implementation for:
1) [TransBTS: Multimodal Brain Tumor Segmentation Using Transformer](https://arxiv.org/abs/2103.04430).
2) [TransBTSV2: Towards Better and More Efficient Volumetric Segmentation of Medical Images](https://arxiv.org/abs/2201.12785).
The details of the our TransBTS and TransBTSV2 can be found at the models directory ([TransBTS](https://github.com/Wenxuan-1119/TransBTS/tree/main/models/TransBTS) and [TransBTSV2](https://github.com/Wenxuan-1119/TransBTS/tree/main/models/TransBTSV2)) in this repo or in the original paper.
## Requirements
- python 3.7
- pytorch 1.6.0
- torchvision 0.7.0
- pickle
- nibabel
## Data Acquisition
- The multimodal brain tumor datasets (**BraTS 2019** & **BraTS 2020**) could be acquired from [here](https://ipp.cbica.upenn.edu/).
- The liver tumor dataset **LiTS 2017** could be acquired from [here](https://competitions.codalab.org/competitions/17094#participate-get-data).
- The kidney tumor dataset **KiTS 2019** could be acquired from [here](https://kits19.grand-challenge.org/data/).
## Data Preprocess (BraTS 2019 & BraTS 2020)
After downloading the dataset from [here](https://ipp.cbica.upenn.edu/), data preprocessing is needed which is to convert the .nii files as .pkl files and realize data normalization.
`python3 preprocess.py`
## Training
Run the training script on BraTS dataset. Distributed training is available for training the proposed TransBTS, where --nproc_per_node decides the numer of gpus and --master_port implys the port number.
`python3 -m torch.distributed.launch --nproc_per_node=4 --master_port 20003 train.py`
## Testing
If you want to test the model which has been trained on the BraTS dataset, run the testing script as following.
`python3 test.py`
After the testing process stops, you can upload the submission file to [here](https://ipp.cbica.upenn.edu/) for the final Dice_scores.
## Reference
1.[setr-pytorch](https://github.com/gupta-abhay/setr-pytorch)
2.[BraTS2017](https://github.com/MIC-DKFZ/BraTS2017)
没有合适的资源?快使用搜索试试~ 我知道了~
温馨提示
该项目利用Transformer深度学习模型进行医学CT和MRI图像的分割。数据集包含大量经过专业医生标注的CT和MRI图像,确保了图像分割的准确性。环境搭建说明详细介绍了所需的软件库、框架版本和硬件配置,帮助用户快速构建实验环境。通过训练和测试,该模型能够准确地从CT和MRI图像中分割出感兴趣的区域,如肿瘤、病变等,为医学研究和临床诊断提供了有力的支持。该项目的成功实施有望提高医学图像分割的效率和准确性,为医生和患者带来更好的医疗服务。
资源推荐
资源详情
资源评论
收起资源包目录
基于transformer的医学CT 和 MRI图像分割内含数据集和环境搭建说明.zip (25个子文件)
data
preprocess.py 3KB
valid.txt 3KB
BraTS.py 5KB
train.txt 8KB
predict.py 9KB
utils
__init__.py 2B
tools.py 207B
figures
qualitative_comparison_TransBTSV2.png 18.61MB
architecture_TransBTSV2.png 6.78MB
architecture_TransBTS.png 6.89MB
quantitative_comparison_TransBTS.png 6.56MB
qualitative_comparison_TransBTS.png 13.3MB
quantitative_comparison_TransBTSV2.png 19.01MB
models
TransBTS
IntmdSequential.py 555B
TransBTS_downsample8x_skipconnection.py 11KB
Transformer.py 3KB
Unet_skipconnection.py 4KB
PositionalEncoding.py 1KB
README.md 1KB
criterions.py 5KB
TransBTSV2
README.md 1KB
README.md 456B
train.py 10KB
test.py 4KB
README.md 2KB
共 25 条
- 1
资源评论
小码蚁.
- 粉丝: 2525
- 资源: 4057
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