A deep learning approach to improve detectionandclassificationofbraintumorinmriimages

Abstract

Brain tumor is considered as one of the most lethal state in cancer where medicalimagingplaysacriticalroleindiagnosis.Inthecurrentscenario,therearevariousradiological process being carried out towards medical image where Magnetic ResonanceImaging (MRI) is considered to offer appropriate localization of the tumor in the brain image.Anextensivereviewofliteratureassociatedwithvariousstandardapproacheshasbeenstudied towards addressing detection process of brain tumor. This approach has advantagesand also limitations in the implementation. However, the common fact observed in all theexisting approaches is that the brain MRI image is directly used in certain detection algorithmwithout considering the fact that there could be possibilities of certain artifacts within MRI.This also results in outliers when various analytical schemes are applied over brain MRIimage in order to perform detection. This is the prime research challenge that has not beenaddressedin existingsystemandhenceitisconsidered inproposedstudy. newline