$\text{SAM-FGF}$

$\text{SAM-FGF}$ is a fine-tuning model based on SAMMed2d, mainly designed for vascular segmentation.

Installation Steps

You can follow the steps below to install:

1. Clone the project to your local machine using the following command:

2. Navigate to the project directory:

   cd SAM-FGF
   

3. Create and activate a virtual environment (optional but recommended):

   conda create -n samfgf python=3.9
   

4. Install all the required dependencies:

   pip install -r requirements.txt
   pip install surface-distance/
   

Usage

Here are some common usage examples: Before that, you need to download the FIVES dataset, which is a high-quality dataset for retinal vessel segmentation.

wget https://s3-eu-west-1.amazonaws.com/pfigshare-u-files/34969398/FIVESAFundusImageDatasetforAIbasedVesselSegmentation.rar?X-Amz-Algorithm=AWS4-HMAC-SHA256&X-Amz-Credential=AKIAIYCQYOYV5JSSROOA/20240619/eu-west-1/s3/aws4_request&X-Amz-Date=20240619T120301Z&X-Amz-Expires=10&X-Amz-SignedHeaders=host&X-Amz-Signature=85f6727c5ddf36ad84c7ae51c69d3341b945d87a1b22e0bb562f77399d74b8aa

Or you can download it from https://figshare.com/articles/figure/FIVES_A_Fundus_Image_Dataset_for_AI-based_Vessel_Segmentation/19688169/1 You may need to pre crop the image into a small image of 256 * 256 size using data_pre.py . Then, you need a pre training weight from SAMMed2d or SAM-vit-b SAMMed2d: download from https://drive.google.com/file/d/1ARiB5RkSsWmAB_8mqWnwDF8ZKTtFwsjl/view?usp=drive_link sam-vit-b:

wget https://dl.fbaipublicfiles.com/segment_anything/sam_vit_b_01ec64.pth

or download from https://github.com/facebookresearch/segment-anything

• Example 1: Run the training script

  python train_without_prompt.py --run_name fives --epochs 60 --batch_size 32 --resume pretrain_model/sam-med2d_b.pth
  

• Example 2: Perform inference using a pre-trained model

  python predict.py --sam_checkpoint ./workdir/models/fives/epoch60_sam.pth
  

Feel free to modify and adjust these examples according to your specific task and requirements.

Contributing Guidelines

If you would like to contribute to this project, please follow these steps:

1. Fork the project and make your modifications.

2. Submit a Pull Request to submit your changes to our repository.

3. We will review your Pull Request and merge appropriate changes.

Copyright and License

This project is licensed under the Apache License. For more details, please refer to the LICENSE.