README.md

metadata

pipeline_tag: mask-generation

SAM-VMNet

SAM-VMNet is a project that combines the power of SAM (Segment Anything Model) and VMNet (Vision-based Medical Network) for advanced medical image segmentation. This project is based on the research paper available at arXiv:2406.00492.

Table of Contents

• Installation
• Usage
• Pre-trained Weights
• License
• Citation

Installation

To set up the project, follow these steps:

1. Clone the repository:

   git clone https://github.com/qimingfan10/SAM-VMNet.git
   cd SAM-VMNet
   

2. Install the required dependencies:

   pip install -r requirements.txt
   

Usage

After installing the dependencies, you can proceed with the following steps:

1. Download the pre-trained weights:

   • Download the following files from Google Drive and place them in the SAM-VMNet/VM-UNet/pre_trained_weights directory:

     • vmamba_small_e238_ema.pth
     • best-epoch142-loss0.3230.pth
     • best-epoch142-loss0.3488.pth

   • Download the following file from Google Drive and place it in the SAM-VMNet/MedSAM-main/work_dir/MedSAM directory:

     • MedSAM_model.pth

2. Run the project:

   • Follow the instructions in the project's source code to run the segmentation tasks.

Pre-trained Weights

The pre-trained weights are essential for the project to function correctly. Ensure that you download and place them in the correct directories as specified in the Usage section.

License

This project is licensed under the MIT License. See the LICENSE file for more details.

Citation

If you use this project in your research, please cite the following paper:

@article{author2024samvmnet,
  title={SAM-VMNet: Advanced Medical Image Segmentation},
  author={Author, First and Author, Second},
  journal={arXiv preprint arXiv:2406.00492},
  year={2024}
}

---

For any questions or issues, please open an issue on this repository.