Tumor segmentation from brain mrimages using u-net & resnet : A comparative study

Abstract

Brain tumors are abnormal growth of cells that need to be detected early for treatment. Magnetic Resonance Imaging (MRI) is a routinely utilized procedure to take brain tumor images. Manual segmentation of tumors is a crucial task and laborious. There is a need for an automated system for segmentation and classification for tumor surgery and medical treatments. This work suggests an efficient brain tumor segmentation and classification based on deep learning techniques. In this thesis work, we proposed a comparative analysis between two brain tumor MRI image segmentation methods U-Net and ResNet. We trained both U-Net and ResNet models using BraTs data and compared the outputs to find out which model works best and gives better accuracy, less loss. In conclusion, this thesis aims to explore the potential of U-Net and ResNet for brain tumor segmentation in medical imaging data. By investigating the effectiveness of U-Net and ResNet architecture, training and evaluating the models, and comparing it with each other, this research seeks to contribute to the field of neuroimaging and facilitate accurate and efficient brain tumor diagnosis and treatment planning. The Ultimate goal is to provide doctors with a reliable tool for improved patient care and management.