SAM Guided Task-Specific Enhanced Nuclei Segmentation in Digital Pathology

Official implementation of the paper published at MICCAI 2024

Report Bug

Qualitative Performance Comparison

Quantitative Performance Comparison

---

Setup

1. Clone the repository:

   git clone https://github.com/CVPR-KIT/SAM-Guided-Enhanced-Nuclei-Segmentation.git

2. Create a new environment and install the required packages:

   pip install -r requirements.txt

Dataset

Three datasets are used in the experiments.

1. CryoNuSeg
2. NuInsSeg
3. CoNIC

The datasets are chosen for their diversity in staining methods and tissue complexity. CryoNuSeg includes a collection of 30 images, NuInsSeg contains 665 image patches, and CoNIC contains 4981 images in the dataset.

Directory Configuration

The datasets can be organised using organize_dataset.py and followed up with genAugmentation.py to generate augmnetations. the code file genSAMEncoding.py can be used to obtain SAM encodings of the images that is to be used during the experiments. The final structure of dataset directory configuration should look something like the following.

NuInsSeg/
  ├── test/
  │   ├── 0.png (input image)
  │   └── 0_label.png (groundtruth image)
  │   └── 0_en.pt (SAM encoding)
  ├── train/
  │   ├── 0.png
  │   └── 0_label.png
  │   └── 0_en.pt 
  └── val/
  │   ├── 0.png
  │   └── 0_label.png
  │   └── 0_en.pt 

Each of the directory has indexed files. each index has 3 corresponding parts - an input image, a label image and SAM encoding.

Experiment Setup

All modifiable parameters related to the experiment and augmentation are present in the config files. Set up all parameters here before proceeding.

Train Model

After checking the dataset information in the config file, run:

python main.py --config config.sys |& tee log/experiments/log-02-15.txt

Inference / Testing

For testing or inferencing images, ensure they are in the correct format and directory information is added correctly in the config.sys file:

python train_test.py --expt_dir <path to experiment directory>

Contributing

Contributions are what make the open source community such an amazing place to learn, inspire, and create. Any contributions you make are greatly appreciated.

• If you have suggestions for adding or removing projects, feel free to open an issue to discuss it, or directly create a pull request after editing the README.md file with necessary changes.
• Please make sure you check your spelling and grammar.
• Create individual PRs for each suggestion.

Creating A Pull Request

1. Fork the Project
2. Create your Feature Branch (git checkout -b feature/newFeature)
3. Commit your Changes (git commit -m 'Added some new feature')
4. Push to the Branch (git push origin feature/newFeature)
5. Open a Pull Request

License

Distributed under the MIT License. See LICENSE for more information.

Authors

• Bishal Ranjan Swain - PhD Candidate at Kumoh National Institute of Technology

Acknowledgements

• Prof. Jaepil Ko