Medical Segmentation Benchmark

Description

This Project is base on Pytorch and Pytorch-Lightning and focus on Medical Image Segmentation

How to run

First, install dependencies

# clone project   
git clone https://github.com/t110368027/Medical_Segmentation_Benchmark

# install project   
cd Medical_Segmentation_Benchmark
pip install -e .   
pip install -r requirements.txt

Next, navigate to any file and run it.

# run module   
python train.py -m MODEL -dn DATASET_NAME
## MODEL UNet, NestedUNet, UNet_3Plus, AttU_Net, R2AttU_Net, and so on
## DATASET_NAME CHUAC, DCA1, STARE, CHASEDB1

Prepare dataset

Before run this project you need to download the dataset :

./Medical_Segmentation_Benchmark/
├── data/
│   ├─angiography
│   │  ├─Hemotool
│   │  │  └─angio1.png_mask.png  
│   │  ├─Original
│   │  │  └─1.png
│   │  └─Photoshop
│   │     └─angio1ok.png
│   ├─CHASE_DB1
│   │  ├─Images
│   │  │  └─Image_01L.jpg
│   │  └─Masks
│   │     └─Image_01L_1stHO.png
│   ├─DB_Angiograms_134
│   │  ├─Database_134_Angiograms
│   │  │  ├─1.pgm
│   │  │  ├─1_gt.pgm
│   ├─STARE
│   │  ├─imgs
│   │  │  └─im0001.png
│   │  └─label-ah
│         └─im0001.ah.png
├── datasets/
│   ├─CHASEDB1
│   │  └─set.npz
│   ├─CHUAC
│   │  └─set.npz
│   ├─DCA1
│   │  └─set.npz
│   └─STARE
│      └─set.npz
├── README.md
└── train.py

Choose a path to create a folder with the dataset name and download datasets

Follow data structure tree before.

Then run preprocess.py convert .jpg, .png or .pgm to numpy array and save it to .npz file

# run module   
python preprocess.py -dp DATASET_PATH -dn DATASET_NAME

## DATASET_PATH ./data
## DATASET_NAME CHUAC, DCA1, STARE, CHASEDB1

Citation

@misc{Medical_Segmentation_Benchmark,
  author = {Jia-Ming Hou},
  title = {{Medical_Segmentation_Benchmark}},
  year = {2023},
  publisher = {GitHub},
  journal = {GitHub repository},
  howpublished = {\url{https://github.com/t110368027/Medical_Segmentation_Benchmark}},
  version = {0.0.1}, 
}

License

Project is distributed under Apache 2.0