Brain Tumor Segmentation Using Volumetric Image Data

Abstract

The number of patients infected with brain tumours rises annually. Inappropriate cell proliferation leads to tumour development. Brain tumours, while mostly malignant, can also be benign (cancerous). It is also possible to classify them as either primary or secondary. Primary tumours develop in the brain and metastasize throughout the body, while secondary tumours have their roots in other organs. The degree of abnormalities in the brain tissue is used to assign one of four grades to the tumours. The standard procedure for diagnosing a brain tumour involves a medical professional looking at MRI scans and making a decision. Obtaining satisfactory accuracy takes time and knowledge from several different individuals. Recently In the field of image classification, deep neural networks have recently become increasingly popular. However, the vast majority of studies on brain tumour segmentation have only used 2D images. A volumetric analysis of MRI scans is crucial for the rapid detection of brain tumours. A brain tumour segmentation utilising U net architecture is the focus of this study. The model examines the Brats dataset, which is available to the public, to draw conclusions about 3D photographs. The findings demonstrate the model's high level of performance.