https://github.com/CUHK-AIM-Group/UN-SAM

https://github.com/CUHK-AIM-Group/UN-SAM

Last synced: about 1 month ago
JSON representation

• Host: GitHub
• URL: https://github.com/CUHK-AIM-Group/UN-SAM
• Owner: CUHK-AIM-Group
• License: mit
• Created: 2024-02-25T14:47:43.000Z (11 months ago)
• Default Branch: main
• Last Pushed: 2024-06-28T09:12:31.000Z (7 months ago)
• Last Synced: 2024-06-29T09:52:15.358Z (7 months ago)
• Language: Python
• Size: 67.4 KB
• Stars: 34
• Watchers: 2
• Forks: 3
• Open Issues: 1
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

• Awesome-Segment-Anything - [code

README

        
# Universal Prompt-Free Segmentation for Generalized Nucleus Images (UN-SAM)

This repository is an official PyTorch implementation of the paper **"UN-SAM: Universal Prompt-Free Segmentation for Generalized Nucleus Images"** [[paper]()] submitted to IEEE Transactions on Medical Imaging.

## Dependencies

* Python 3.10

* PyTorch >= 1.10.0

* albumentations 1.5.2

* monai 1.3.0

* pytorch_lightning 1.1.0

## Code

Clone this repository into any place you want.

```bash

git clone https://github.com/CUHK-AIM-Group/UNSAM.git

cd UNSAM

mkdir data; mkdir pretrain;

```

## Quickstart 

* Train the UN-SAM with the default settings:

```python

python train.py --dataset data/$YOUR DATASET NAME$ --sam_pretrain pretrain/$SAM CHECKPOINT$

```

## Cite

If you find our work useful in your research or publication, please cite our work:

```

@article{chen2024sam,

  title={UN-SAM: Universal Prompt-Free Segmentation for Generalized Nuclei Images},

  author={Chen, Zhen and Xu, Qing and Liu, Xinyu and Yuan, Yixuan},

  journal={arXiv preprint arXiv:2402.16663},

  year={2024}

}

```

## Acknowledgements

* [SAM](https://github.com/facebookresearch/segment-anything)