Open AccessComparative analysis of glioma tumor in brain using machine learning and deep learning techniques
Author(s) -
J. Sangeetha,
D. Vaishnavi,
J. Premalatha
Publication year - 2021
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1767/1/012040
Subject(s) - artificial intelligence , support vector machine , convolutional neural network , computer science , random forest , abstraction , glioma , deep learning , machine learning , brain tumor , magnetic resonance imaging , artificial neural network , pattern recognition (psychology) , pathology , medicine , radiology , cancer research , philosophy , epistemology
Brain tumour is one of the major causes of the increasing mortality rate. There is an emerging need for precise diagnosis from the pathology, as human prediction is prone to errors. Brain pathology is captured using Magnetic Resonance Imaging (MRI) which gives high quality images of the blood vessels. These MRI images can be used to extract features and predict the tumor. The extracted features can be classified using Machine leaning (ML) algorithms. In this advanced and emerging deep learning era, images can be fed directly to the prediction system and the system itself extracts features at higher level of abstraction and classifies the images. This article presents a comparative analysis for classifying the MRI images of Glioma (a type of brain tumor) and healthy brain using convolutional neural networks (CNN) and ML algorithms like support vector machines (SVM), random forests, and bagging. CNN and the ML algorithms were implemented on BRATS 2013 challenge dataset and it is found that CNN has achieved highest accuracy of 95%.