Project Brain Segmentation

📁 Data Format

Folder Structure

data_folder/
├── data/               # Contains .h5 data files
├── train.list          # List of training samples
├── val.list            # List of validation samples
└── test.list           # List of test samples

H5 File Structure

Each .h5 file in the data/ folder contains two keys:

• image — the input image array
• label — the corresponding segmentation mask array

Example of how data is loaded:

case = self.sample_list[idx]
if self.split == "train":
    h5f = h5py.File(self._base_dir + "/data/{}".format(case), 'r')
else:
    h5f = h5py.File(self._base_dir + "/data/{}".format(case), 'r')

image = h5f['image'][:]
label = h5f['label'][:]

---

🚀 Run Code

Training & Testing Scripts

The project includes multiple paired training and testing scripts. Configuration parameters can be modified directly inside each script.

Script	Description
train_BCP.py	Training script for BCP method
test_BCP.py	Testing script for BCP method

Note: Additional train_*.py / test_*.py pairs may be available for other methods. Follow the same pattern to run them.

Running Training

python train_BCP.py

Running Testing

python test_BCP.py

Configuration

All hyperparameters and settings (e.g., dataset path, number of epochs, batch size, learning rate) can be configured by editing the config section at the top of each script before running.