[Preprint article] [Dataset & Pretrained models][Poster]
Pulmonary diseases, particularly lung cancer, have become major challenges for public health. Thoracic computed tomography (CT) plays a vital role in the early detection of lung cancer. With the rapid development of artificial intelligence (AI), AI-assisted medical imaging analysis has emerged, demonstrating remarkable performance across diverse medical applications. However, the costly annotation process and privacy concerns limit the construction of large-scale medical datasets, hampering the further application of AI in healthcare. To address the data scarcity in medical imaging, we propose Lung-DDPM, a thoracic CT image synthesis approach that effectively generates high-fidelity 3D synthetic CT images. These images prove helpful in downstream lung nodule segmentation tasks. Our method is based on semantic layout-guided denoising diffusion probabilistic models (DDPM), enabling anatomically reasonable, seamless, and consistent sample generation even from incomplete semantic layouts. Our results suggest that the proposed method outperforms other state-of-the-art (SOTA) generative models in image quality evaluation and downstream lung nodule segmentation tasks. Specifically, Lung-DDPM achieved superior performance on our large validation cohort, with a Fréchet inception distance (FID) of 0.0047, maximum mean discrepancy (MMD) of 0.0070, and mean squared error (MSE) of 0.0024. These results were 7.4×, 3.1×, and 29.5× better than the second-best competitors, respectively. Furthermore, the lung nodule segmentation model, trained on a dataset combining real and Lung-DDPM-generated synthetic samples, attained a dice coefficient (Dice) of 0.3914 and sensitivity of 0.4393. This represents 8.8% and 18.6% improvements in DICE and sensitivity compared to the model trained solely on real samples. The experimental results highlight Lung-DDPM’s potential for a broader range of medical imaging applications, such as general tumor segmentation, caner survival estimation and risk prediction.
|
Ground truth |
Semantic layout |
Med-DDPM |
Make-A-Volume |
GEM-3D |
Lung-DDPM |
|
|
Case 1 |
|
|
|
|
|
|
|
Case 2 |
|
|
|
|
|
|
conda env create -n lung-ddpm python=3.9
conda activate lung-ddpm
conda install pytorch==1.13.1 torchvision==0.14.1 torchaudio==0.13.1 pytorch-cuda=11.6 -c pytorch -c nvidia
pip install -r requirements.txt
We prepared three sample cases at GoogleDrive, and you can download to 'data/LIDC-IDRI/' folder. Then you can easily use a command to verify the training and sampling function.
Also, you are welcome to specify '--ct_path' and '--mask_path' parameters for your own dataset, please follow the above three cases for data formats.
Download the pretrained model at GoogleDrive. And put it at the 'checkpoints' folder in the root directory. The directory structure should look similar as follows:
Lung-DDPM/
│── data/
│ ├── LIDC-IDRI/ # Prepared dataset
│ │ ├── CT/
│ │ │ ├── LIDC-IDRI-0001.nii.gz
...
│ │ ├── SEG/
│ │ │ ├── LIDC-IDRI-0001.nii.gz
...
│ ...
│── checkpoints/ # Pretrained models
│ ├── Lung-DDPM-LIDC-IDRI-100000-steps.pt
│── train.py
│── sample.py
│── README.md
...
python train.py --ct_path "data/LIDC-IDRI/CT" --mask_path "data/LIDC-IDRI/SEG" --batchsize 1 --epochs 100000 --input_size 128 --depth_size 128 --num_class_labels 3 --num_channels 64 --num_res_blocks 1 --timesteps 250 --save_and_sample_every 2
python sample.py --ct_path "data/LIDC-IDRI/CT" --mask_path "data/LIDC-IDRI/SEG" --sample_path "data/LIDC-IDRI/SAMPLE" --vis_path "data/LIDC-IDRI/VIS" --input_size 128 --depth_size 128 --num_class_labels 3 --num_channels 64 --num_res_blocks 1 --timesteps 250 --batchsize 1 --num_samples 1 --num_class_labels 3 --timesteps 250 --weightfile "checkpoints/Lung-DDPM-LIDC-IDRI-100000-steps.pt" --blend_from 250
After the sampling process done, the sampling results will be available at 'sample_path' and the visualization results will be available at the 'vis_path'.
- Release training and sampling code.
- Release pretrained models.
- Release the LIDC-IDRI patient list.
- A synthetic dataset based on LIDC-IDRI.
- HuggingFace online demo.
- Lung-DDPM+: we are working on an efficient version of Lung-DDPM. The code will be available soon. Stay tuned!
If Lung-DDPM contributes to your research, please cite as follows:
@misc{jiang2025lungddpmsemanticlayoutguideddiffusion,
title={Lung-DDPM: Semantic Layout-guided Diffusion Models for Thoracic CT Image Synthesis},
author={Yifan Jiang and Yannick Lemaréchal and Josée Bafaro and Jessica Abi-Rjeile and Philippe Joubert and Philippe Després and Venkata Manem},
year={2025},
eprint={2502.15204},
archivePrefix={arXiv},
primaryClass={eess.IV},
url={https://arxiv.org/abs/2502.15204},
}
Lung-DDPM is developed based on the following code repositories:
We are very grateful for their contributions to the community.