UDTdiffusion: A Pseudo-Condition Guided Adversarial Fast Diffusion Model for Unpaired Medical Image Translation
After the article is accepted, we will provide trained .pth weight files for researchers to reproduce the test results
python>=3.6.9
torch>=1.7.1
torchvision>=0.8.2
cuda=>11.2
ninja
python3.x-dev (apt install, x should match your python3 version, ex: 3.8)
git clone https://github.com/uuuaziQAQ/UDTdiffusion
cd UDTdiffusion
input_path/
── data_train_contrast1.mat
── data_train_contrast2.mat
── data_val_contrast1.mat
── data_val_contrast2.mat
── data_test_contrast1.mat
── data_test_contrast2.mat
where .mat files has shape of (#images, width, height) and image values are between 0 and 1.0.
You can use file make_mat.m to convert the .jpg images to .mat format
IXI dataset：https://brain-development.org/ixi-dataset/
The Gold Atlas Male Pelvis Dataset：T. Nyholm et al., “MR and CT data with multiobserver delineations of organs in the pelvic area—part of the gold atlas project,” Med. Phys., vol. 45, no. 3, pp. 1295–1300, 2018.
The breast dataset: If researchers need to use it for scientific research, you can contact our email [email protected] and provide it
python test.py --image_size 256 --exp UDT --num_channels 2 --num_channels_dae 64 --ch_mult 1 1 2 2 4 4 --num_timesteps 4 --num_res_blocks 2 --batch_size 1 --embedding_type positional --z_emb_dim 256 --contrast1 contrast1 --contrast2 contrast2 --which_epoch 100 --gpu_chose 1 --input_path breast --output_path results