This project is the implementation of our method(Lequan Yu, Caizi Li) for MICCAI Grand Challenge on 6-month Infant Brain MRI Segmentation from Multiple Sites(iSeg-2019). Our Code is based on 3D_DenseSeg and ADVENT.
- PyTorch 1.2
- Python 3.5
- Ubuntu 16.04
- Cuda 10.0
- PyCharm 2019.3.3 (Community Edition)
- batchgenerators
- GeForce RTX 2080Ti
-
Step 1: Change the root directory 'PROJECT_ROOT' into your owns in Config.config
-
Step 2: Put the training data, validation data and testing data in 'PROJECT_ROOT/Dataset/src/iSeg-2019-Training', 'PROJECT_ROOT/Dataset/src/iSeg-2019-Validation' and 'PROJECT_ROOT/Dataset/src/iSeg-2019-Testing', respectively.
-
Step 3: cd 'Data_preprocessing', generate hdf5 files for data of 'Step 2' by 'prepare_hdf5_cutedge.py', 'prepare_hdf5_cutedge_valdata.py' and 'prepare_hdf5_cutedge_testdata.py'. The hdf5 files for training, validation and testing will be found in directory 'PROJECT_ROOT/Dataset/hdf5_iseg_data', 'PROJECT_ROOT/Dataset/hdf5_iseg_val_data' and 'PROJECT_ROOT/Dataset/hdf5_iseg_test_data', respectively.
-
Step 4: cd 'Main', run 'train.py' and 'test.py' for training and testing.