Magnetic Resonance Imaging

6: CNN-based segmentation of brain tumor from T2-weighted MRI slices

Abstract

The brain is a major organ in human physiology, and plays a main function related to sensory information handing and decision-making. Deficiencies in the brain cause various problems including temporary/permanent disability and death. Brain tumor is one of the brain abnormalities and based on its orientation and severity, it can be classified as glioma (low-grade glioma and high-grade glioma) and glioblastoma multiforme (GBM). This work aims to utilize a convolutional neural network (CNN)-based automated scheme to obtain the cancer fragments from 2D MRI slices. In this work, the well-known pretrained CNN methods, such as UNet, SegNet, VGG-UNet, and VGG-SegNet are ...