A sophisticated convolutional neural network model for brain tumor classification

Magnetic Resonance Imaging(MRI) is one of the commonly used medical imaging modality that provides informative data for brain tumor diagnosis other than Computed Tomography(CT). A key challenge when a physician studies the MRI data is the time and effort he has to put in diagnosing the tumors. The objective of this research is to recognize the tumor type when a collection of MRI images of a patient is given. To achieve this goal, a deep learning algorithm is developed using Convolutional Neural Networks(CNNs). Nowadays, most of image classification problems use CNNs as they deliver higher precision and accuracy compared to other existing algorithms. Here, a sophisticated CNN model is developed, trained using cross validation and tested on brain MRI images obtained from open databases. The performance of the proposed model is promising.