Skip to content
master
Switch branches/tags
Code

Latest commit

 

Git stats

Files

Permalink
Failed to load latest commit information.
Type
Name
Latest commit message
Commit time
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Generation of Virtual Dual Energy Images from Standard Single-Shot Radiographs using Multi-scale and Conditional Adversarial Network

Bo Zhou, Xunyu Lin, Brendan Eck, Jun Hou, David L. Wilson

Asian Conference on Computer Vision (ACCV), 2018

[Paper]

This repository contains the PyTorch implementation of MCA-Net for DE bone image generation.
Pre-trained model is available upon request via:
https://drive.google.com/file/d/1agAASv1B5Uecxh9uyt4q2i-6NTDRLPYc/view?usp=sharing

The code and model are for research use only.

We provide an example case in the './example_data/'

Citation

If you use this code for your research or project, please cite:

@inproceedings{zhou2018generation,
  title={Generation of virtual dual energy images from standard single-shot radiographs using multi-scale and conditional adversarial network},
  author={Zhou, Bo and Lin, Xunyu and Eck, Brendan and Hou, Jun and Wilson, David},
  booktitle={Asian Conference on Computer Vision},
  pages={298--313},
  year={2018},
  organization={Springer}
}

Environment and Dependencies

Requirements:

  • Python 3.7
  • Pytorch 0.4.1
  • scipy
  • scikit-image
  • opencv-python
  • tqdm

Our code has been tested with Python 3.7, Pytorch 0.4.1, CUDA 10.0 on Ubuntu 18.04.

Dataset Setup

.
example_data
├── Train                   # contain training files (IB:bone image / IS:soft-tissue image / IH:standard chest x-ray image)
│   ├── IB
│   │   ├── IB_1.png         
│   │   ├── IB_2.png 
│   │   ├── ...         
│   │   └── IB_N.png 
│   │   
│   ├── IS
│   │   ├── IS_1.png         
│   │   ├── IS_2.png 
│   │   ├── ...         
│   │   └── IS_N.png 
│   │   
│   ├── IH
│   │   ├── IH_1.png         
│   │   ├── IH_2.png 
│   │   ├── ...         
│   │   └── IH_N.png 
│   └── ...
│
│
├── Test                    # contain test files (IB:bone image / IS:soft-tissue image / IH:standard chest x-ray image)
│   ├── IB
│   │   ├── IB_1.png         
│   │   ├── IB_2.png 
│   │   ├── ...         
│   │   └── IB_N.png 
│   │   
│   ├── IS
│   │   ├── IS_1.png         
│   │   ├── IS_2.png 
│   │   ├── ...         
│   │   └── IS_N.png 
│   │   
│   ├── IH
│   │   ├── IH_1.png         
│   │   ├── IH_2.png 
│   │   ├── ...         
│   │   └── IH_N.png 
│   └── ...
│            
└── ...

Each .png is an image data and intensity normalized to between 0~255. IB_N.png / IS_N.png / IH_N.png should contain paired imaging data for the same patient.

To Run Our Code

  • Train the model
python train.py --experiment_name 'train_bone_msunet' --model_type 'model_bone' --dataset 'DE' --data_root './example_data/' --net_G 'msunet' --net_D 'patchGAN' --wr_recon 50 --batch_size 2 --lr 1e-4 --AUG

where
--experiment_name provides the experiment name for the current run, and save all the corresponding results under the experiment_name's folder.
--data_root provides the data folder directory (with structure illustrated above).
--AUG adds for using data augmentation option (rotation, random cropping, scaling).
Other hyperparameters can be adjusted in the code as well.

  • Test the model
python test.py --resume './output/train_bone_msunet/checkpoints/model_best.pt' --experiment_name 'test_bone_msunet' --model_type 'model_bone' --dataset 'DE' --data_root './example_data/' --net_G 'msunet' --net_D 'patchGAN'

where
--resume defines which checkpoint for testing and evaluation. The 'model_best.pt' is available upon request.
The test will output an eval.mat containing model's input and prediction for evaluation in the '--experiment_name' folder.

Sample training/test scripts are provided under './scripts/' and can be directly executed.

Contact

If you have any question, please file an issue or contact the author:

Bo Zhou: bo.zhou@yale.edu

About

Generation of Virtual Dual-Energy Images from Standard Single-Shot Radiographs using Multi-scale and Conditional Adversarial Network (ACCV 2018)

Resources

Releases

No releases published

Packages

No packages published