SpineFM: Leveraging Foundation Models for Automatic Spine X-ray Segmentation

https://arxiv.org/abs/2411.00326

Abstract

This paper introduces SpineFM, a novel pipeline that achieves state-of-the-art performance in the automatic segmentation and identification of vertebral bodies in cervical and lumbar spine radiographs. SpineFM leverages the regular geometry of the spine, employing a novel inductive process to sequentially infer the location of each vertebra along the spinal column. Vertebrae are segmented using Medical-SAM-Adaptor, a robust foundation model that diverges from commonly used CNN-based models.

Installation

Requirements

• Python 3.10.8

• Required libraries/packages:

  pip install -r requirements.txt

• Model Weights

• Datasets

Usage

Running the Code

To reproduce the results from the paper, you should:

• Download either dataset (although currently NHANES II website is down)
• Download the corresponding model weights
• Ensure that the utils.py get_model() function weight file names match with your own
• Run the code:

python main.py "output_directory" "weights_path" "dataset*" "data_path"

*either NHANES II or CSXA

Model Training

If you want to replicate this with a new dataset then I recommend getting in touch with me and I will try help. To start with, some code is included for training the Mask R-CNN, ResNet and Point_Predictor models, although this hasn't been polished. For fine-tuning of the Medical-SAM-Adaptor see the original repo. I can provide extra details of the training process for this model if needed.

Citation

If you use this code in your research, please cite our paper:

@article{simons2025spinefmleveragingfoundationmodels,
      title={SpineFM: Leveraging Foundation Models for Automatic Spine X-ray Segmentation}, 
      author={Samuel J. Simons and Bartłomiej W. Papież},
      year={2025},
      eprint={2411.00326},
      archivePrefix={arXiv},
      primaryClass={eess.IV},
      url={https://arxiv.org/abs/2411.00326}, 
}

License

This code is released under the GPL-3.0 License. See the LICENSE.txt file for details.