Code of High-Resolution Chest X-ray Bone Suppression Using Unpaired CT Structural Priors
Python==3.6
Torch==1.1.0
Scipy ==1.1.0 if your Scipy>=1.3.0, you will meet “ImportError: cannot import name 'imresize'”
There three steps of our work: 1. cycleganxray2gradient, 2. Pytorch-UNet-master_trainbonemask and 3. histogram_match.
Please download our whole project code HERE code: 6orv (including weight, test imges) instead of the code in this github.
Pleas feel free to contact me if you have any problem: han.li[at]miracle.ict.ac.cn / 137412918[at]qq.com
cd 1cycleganxray2gradient
python test.py --dataroot ./datasets --name maps_cyclegan4 --model cycle_gan --epoch 110
Input data location: ‘./datasets/testA’
Output data location: ’ ./results/maps_cyclegan4/test_110/images’
cd ..
cd 2Pytorch-UNet-master_trainbonemask
python newpredict.py -c ./checkpoints/CP50.pth
Input data location: Output data location in step1
Output bone data location: ’ ./bone ’
Output lung mask location: ’ ./lung ’
cd ..
cd 3histogram_match
python bonemask2result.py
input data location: ’ ./dataset'’
final result location : ‘./result’