https://github.com/dong845/MCP-MedSAM

https://github.com/dong845/MCP-MedSAM

Last synced: 2 months ago
JSON representation

• Host: GitHub
• URL: https://github.com/dong845/MCP-MedSAM
• Owner: dong845
• License: mit
• Created: 2024-11-29T19:42:41.000Z (2 months ago)
• Default Branch: main
• Last Pushed: 2024-12-10T09:23:41.000Z (2 months ago)
• Last Synced: 2024-12-10T10:27:35.177Z (2 months ago)
• Size: 2.93 KB
• Stars: 1
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

• Awesome-Segment-Anything - [code

README

        
# MCP-MedSAM

Pytorch Implementation of the paper "[MCP-MedSAM: A Powerful Lightweight Medical Segment Anything Model Trained with a Single GPU in Just One Day](https://arxiv.org/abs/2412.05888)". This paper introduces a new variant of MedSAM, integrating a lightweight pre-trained tiny ViT, two novel prompts (modality prompt and content prompt), and a modality-based data sampling strategy. These enhancements enable the model to achieve strong performance without long training time and large GPU resource consumption. The code will come soon.

## Citation

```bash

@misc{lyu2024mcpmedsampowerfullightweightmedical,

      title={MCP-MedSAM: A Powerful Lightweight Medical Segment Anything Model Trained with a Single GPU in Just One Day}, 

      author={Donghang Lyu and Ruochen Gao and Marius Staring},

      year={2024},

      eprint={2412.05888},

      archivePrefix={arXiv},

      primaryClass={cs.CV},

      url={https://arxiv.org/abs/2412.05888}, 

}

```