Comparison of different CNN architectures for brain tumor detection using fMRI

Abstract

Brain is the most vital organ of human body which controls the entire nervous system of human body. In that case, if anything goes wrong inside our brain the entire nervous system gets collapsed. The brain tumors are the most severe and harmful disease, resulting in a very short life expectancy of the affected patient. Thus, ensuring proper treatment at the early stage is the way to provide the quality of life of patients. Detection of brain tumor is a challenging task in the early stages but with the help of modern technology and machine learning algorithms, it has become a matter of great interest. While detecting brain tumor of an affected person we are considering the fMRI data of patient. Our task is to identify whether the tumor is present in patient’s brain or not. Our machine learning algorithm will be convolutional neural network(CNN) that is good enough to generate higher accuracy. We have used some deeper architecture design VGG16 and VGG19 for the better accuracy and comparison purpose. We have done three kinds of classification with these architectures, they are binary classification, five Class Brain Lobe Classification,4 position classification. These different architectures will produce different accuracy level through CNN. Different architectures with different classification will help to find that which one of them meets up the best accuracy level.