learnopencv/Training_3D_U-Net_Brain_Tumor_Seg

Training 3D U-Net for Brain Tumor Segmentation (BraTS-GLI)

This folder contains the Jupyter Notebooks for the LearnOpenCV article - Training 3D U-Net for Brain Tumor Segmentation (BraTS2023-GLI).

We have provided:

• code to prepare dataset (01_Data_Preprocessing.ipynb)
• Model Training Script (02_Training_3D_U-Net-83.59iou.ipynb)
• Video inference code, at the last section of (02_Training_3D_U-Net-83.59iou.ipynb)

Model Download:

• Download trained model from here and place it under model_checkpoint/3D_UNet_Brats2023/version_0/ckpt.tar . One can simply change the relative model path according to your notebook.

Instructions:

• Run the 01_Data_Preprocessing.ipynb to download and preprocess the BraTS2023-GLI Subset from Kaggle.
• After data preprocessing steps, switch to 02_Training_3D_U-Net-83.59iou.ipynb for custom data loader preparation, model definition, training and inference scripts.

You can download the trained weights and code files as well from the below link.

AI Courses by OpenCV

Want to become an expert in AI? AI Courses by OpenCV is a great place to start.