This repository is an implementation of the paper 'Parameterized Gompertz-guided Morphological AutoEncoder for Predicting Pulmonary Nodule Growth', accepted by TMI.

Gompertz network for volume and mass growth prediction on the baseline nodule over time.

Architecture of morphological autoencoder equipped with shape-aware (DS) and texture-aware (DN ) decoders.

Run cd gompertz python main_ddp.py --resume --gpu 1,2 --batch-size 16 --epochs 60 --distributed

cd gmae python main_ddp.py

Please download the trained models form baidu, key is 'qntc', and place the downloaded models according to directory structure.

Run cd gompertz python gompertz/inference.py

cd gmae python inference.py

Details of data processing in STMixer. The paired nodules selected for this project is listed in './data/data.csv' (NLSTt), and please download raw CT scans from NLST

STATISTICS (IN Nn/Ns) OF BENCHMARK SPLITS OF THE NLSTT DATASET AND IN-HOUSE DATASET*.

The center three columns show the predicted texture (left) and shape (right, solid yellow line) of time points Nt by PredNet, NoFoNet, and our GM-AE.

This repository is released under the Apache 2.0 license as found in the LICENSE file.

Please cite GM-AE in your publications if it helps your research.

@article{fang2023parameterized,
  title={Parameterized Gompertz-guided Morphological AutoEncoder for Predicting Pulmonary Nodule Growth},
  author={Fang, Jiansheng and Wang, Jingwen and Li, Anwei and Yan, Yuguang and Liu, Hongbo and Li, Jiajian and Yang, Huifang and Hou, Yonghe and Yang, Xuening and Yang, Ming and others},
  journal={IEEE Transactions on Medical Imaging},
  year={2023},
  publisher={IEEE}
}