Semantic-Segmentation

Dataset BRaTS 21

Monai module SegResNet model

This is an example of semantic segmentation on MRI scan. To run this code a GPU connected environment and about 25 GB of free space is required, the first cell is for all the necessary libraries are needed to run the code, colab usually comes with all the preinstalled libraries with compatible version, using colab is recomended. If this code is run in the local machine version compatibility with different libraries might be an issue. To train the model on entire dataset need a very powerful GPU with around 80GB of RAM to get the best results. The training period may take upto several hours depending on the resources. Cell number 8 download the entire dataset. I trained the model with limited data for resource constrain. Update the 'task' in cell 8 to change any number of test cases.