This repository contains supporting material of the publication Lesion-conditioning of synthetic MRI-derived subtraction-MIPs of the breast using a latent diffusion model.
Citation:
Kapsner, L.A., Folle, L., Hadler, D. et al. Lesion-conditioning of synthetic MRI-derived subtraction-MIPs of the breast using a latent diffusion model. Sci Rep 14, 6391 (2024). https://doi.org/10.1038/s41598-024-56853-1
At first, prepare your data accordingly and adjust the config file as well as the data loader.
To run the code, then execute the following commands:
git clone -b v0.1.0 https://github.com/kapsner/latent-diffusion
cd latent-diffusion
conda env create -f environment.yaml
conda activate ldm
export parameter_grid=./param_grid_230119.csv
export ldm_cfg=./configs/latent-diffusion/dce_mip-vq-seg.yaml
export results_folder=/home/user/development/trainings/diffusionmodels
python automate_learnings.py \
-c $parameter_grid \
-b $ldm_cfg \
-l $results_folder