Learning Interpretable Lesion-Focused Saliency Maps with Context-Aware Diffusion Model for Vertebral Disease Diagnosis

Overview

Environment and Data Preparation

To set up the environment, use the requirements.txt file to prepare the Conda environment. Ensure you have all necessary dependencies installed.

conda env create -f environment.yml
conda activate diff-lfsm

As an example, to prepare the VerTumor600 dataset, convert the MRI vertebrae data from JSON to PNG format using the provided script:

python ./data/VerTumor600/MRI_vertebrae/json_to_png.py

Training and Testing

1. Train Diff-LFSM

To train the Diff-LFSM model, run the following command:

python diff_training_seg_training.py ARG_NUM=1

2. Generate LFSM using the Trained Model

Once the Diff-LFSM model is trained, use it to generate the Lesion-Focused Saliency Maps (LFSM):

python diff_training_seg_training.py ARG_NUM=2

3. Crop Individual Vertebrae

After generating LFSMs, crop out individual vertebrae from the images. For the VerTumor600 dataset, use the script below:

python ./data/VerTumor600/cropped_vertebrae/crop_images.py

4. Train and Test the Classifier

python spine_trans_cls_with_FE.py ARG_NUM=3

Acknowledgement

Thanks to the following works: AnoDDPM, guided-diffusion.