Inference Notebook: https://www.kaggle.com/code/iasonasxrist/mri-brain-tumour-segmentation-with-unet-cnn?scriptVersionId=107332844
Quantitative analysis of brain MRI is routine for many neurological diseases and conditions and relies on accurate segmentation of structures of interest. Deep learning-based segmentation approaches for brain MRI are gaining interest due to their self-learning and generalization ability over large amounts of data.
Here, 2 pre-trained model architectures: UNet and Unet with ResNext backbone were trained and evaluated for Dice scores on Brain MRI dataset obtained from The Cancer Imaging Archive (TCIA).
Dataset: Number/Size of Images : Total : 3929 Training set : 2947 Validation set : 393 Test set : 797(approx)
University of West Attica
Running Pre-installation:Tensorflow,Keras,nibabel,sklearn,numpy,pandas, PIL
Download and unzip the dataset from Kaggle
Results Visualization :
Disclaimer and known issues These codes are implemented in Tensorflow, Pytorch All trainings have been executed into kaggle enviroment due to GPU availability. You can find and fork all my implementation into my kaggle.
![]() |
![]() |
![]() |