Are you sure you want to delete this task? Once this task is deleted, it cannot be recovered.
linyebin123 b0128b7bed | 1 year ago | |
---|---|---|
.. | ||
augmentation | 1 year ago | |
crossvalidation | 1 year ago | |
README.md | 1 year ago | |
autoimplant_amp_config.json | 1 year ago | |
autoimplant_config.json | 1 year ago | |
autoimplantenv.yml | 1 year ago | |
subjects_augmented.json | 1 year ago |
git clone https://github.com/ellisdg/3DUnetCNN.git
cd 3DUnetCNN
export PYTHONPATH=${PWD}:${PYTHONPATH}
cd
into the autoimplant2020
example directory:cd examples/autoimplant2020
https://zenodo.org/record/4270278
python -W ignore::UserWarning:nilearn.image.resampling:273 ../../unet3d/scripts/train.py --config_filename ./autoimplant_config.json --model_filename ./autoimplant_unet3d_fold0.h5 --training_log_filename ./autoimplant_unet3d_fold0_training_log.csv --nthreads <nthreads> --ngpus <ngpus> --fit_gpu_mem <gpu_mem>
<nthreads>
,
<ngpus>
, and
<gpu_mem>
should be set to the number of threads, number of GPUs, and the amount of GPU memory in GB on a single gpu that will be used for training.
-W ignore:UserWarning:nilearn.image.resampling:273
ignores a warning that
nilearn outputs when you resample an image with all zeros and ones using linear or continuous resampling.
Normally, we would not want to resample a labeled image in this way, but in this case we do. The linear
resampling makes the interpolated voxels carry less weight than they would using nearest neighbor.
Note that the training will take a long time. Training on 2 V100 GPUs took 7 days to complete 35 epochs.
python ../../unet3d/scripts/predict.py --output_directory ./predictions/validation/fold0 --config_filename ./crossvalidation/autoimplant_config_fold0.json --model_filename ./autoimplant_unet3d_fold0.h5 --group validation --output_template "predicted_complete_skull.nii.gz" --nthreads <nthreads> --ngpus <ngpus>
<nthreads>
and
<ngpus>
should be set to the number of threads and gpus that you are using.
The predicted complete skull volumes will be in the folder: ./predictions/validation/fold0
See for Configuration.md for tips on how to customize the configuration.
Ellis D.G., Aizenberg M.R. (2020) Deep Learning Using Augmentation via Registration: 1st Place Solution to the AutoImplant 2020 Challenge. In: Li J., Egger J. (eds) Towards the Automatization of Cranial Implant Design in Cranioplasty. AutoImplant 2020. Lecture Notes in Computer Science, vol 12439. Springer, Cham. https://doi.org/10.1007/978-3-030-64327-0_6
用于医学图像分割的Pytorch 3D U-Net卷积神经网络(CNN)
Python
Dear OpenI User
Thank you for your continuous support to the Openl Qizhi Community AI Collaboration Platform. In order to protect your usage rights and ensure network security, we updated the Openl Qizhi Community AI Collaboration Platform Usage Agreement in January 2024. The updated agreement specifies that users are prohibited from using intranet penetration tools. After you click "Agree and continue", you can continue to use our services. Thank you for your cooperation and understanding.
For more agreement content, please refer to the《Openl Qizhi Community AI Collaboration Platform Usage Agreement》