The UNet Model for Multimodal Brain Tumor Classification and Segmentation

MRI manual detection relies on the ability and experience of the physicians. This study presents a brain tumor segmentation and identification system to address this issue and experiments on the BraTS dataset which provides four types of scans, i.e. T1, T2, T1CE, and FAIR, allowing the detection of different class labels and establishing the ground truth for brain tumor segmentation. UNet was used to develop a fully automated approach for the segmentation of glioma on preoperative MRI scans. Images were preprocessed and then classified using the UNet CNN.