Abnormal Tumor Classification of MR Brain Images Using Deep Neural Networks

Abstract

A cell in the brain is a smallest unit of tissue and its abnormal growth leads to brain tumor. This paper presents an effective way of classifying the abnormal MR brain images using customized 10-layer fully connected deep neural network. The process starts with removing of skull region from the brain image using image processing methods and later include pseudo color transformation of the axial T1-weighted MR brain images where the gray scale images are converted into “Lab” color domain. Multiple regions of the brain are labeled which are been trained for the custom designed deep neural network. The classification accuracy of the proposed network attained 98% which is superior than earlier SVM and HMM classifier methods.