BraTS 2023 Adult Glioma

This project folder contains the configuration files, preprocessing, and visualization scripts for the BraTS2023AdultGlioma dataset.

For more information, please refer to https://www.synapse.org/#!Synapse:syn51156910/wiki/.

Related papers:

• https://arxiv.org/pdf/1811.02629.pdf,

• https://arxiv.org/pdf/2305.17033.pdf.

Data need to be downloaded manually due to required registration. Download link: https://www.synapse.org/#!Synapse:syn51156910/wiki/622351.

Note

When running the preprocessing scripts please make sure you have the following packages installed: argparse, json, nibabel, numpy, pathlib, random, tqdm. Those packages are installed by default if atommic is installed.

Visualization An example notebook for visualizing the data is provided in the visualize notebook. You just need to set the path where the dataset is downloaded.

Preprocessing The preprocessing pipeline is implemented in the preprocess_dataset.sh script, consisting of the following steps: 1. Crop to the brain region, as there is a lot of background around the brain resulting is slower training. Important note: the cropping is done only for the training set. 2. Normalize the images to zero mean and unit variance. 3. Updates headers and save to NIfTI format. 4. Split the dataset into training and validation sets. 5. Compute the probabilities for each segmentation class.

Training/Testing

Important

The BraTS2023AdultGlioma dataset is natively supported in atommic. Therefore, you do not need to create a custom dataset class. You just need to set the dataset_format argument in the configuration file to BraTS2023AdultGlioma. For example:

train_ds:
    dataset_format: BraTS2023AdultGlioma

validation_ds:
    dataset_format: BraTS2023AdultGlioma

test_ds:
    dataset_format: BraTS2023AdultGlioma

For training a model, you just need to set up the data and export paths to the configuration file of the model you want to train. In train_ds and validation_ds please set the data_path to the generated json files. In exp_manager please set the exp_dir to the path where you want to save the model checkpoints and tensorboard or wandb logs.

You can train a model with the following command:

atommic run -c /projects/SEG/BraTS2023AdultGlioma/conf/train/{model}.yaml

For testing a model, you just need to set up the data and export paths to the configuration file model you want to test. In checkpoint (line 2) set the path the trained model checkpoint and in test_ds please set the data_path. In exp_manager please set the exp_dir to the path where the predictions and logs will be saved.

You can test a model with the following command:

atommic run -c /projects/SEG/BraTS2023AdultGlioma/conf/test/{model}.yaml

Note: The default logger is tensorboard.