Fusion of brain MRI and CT images and Classification of brain tumor using Machine Learning Algorithm

Abstract

Brain tumor is the most common and destructive disease which reduces the life time of people. The earlier detection of brain tumor plays a most important role for better treatment of the patient. In this paper, a new technique for brain tumor classification using machine learning by fusion of MRI and CT images are proposed. Image fusion is a process of fusing two or more images (i.e. MRI and CT scan images) to obtain a new one which contains more accurate information of the brain than any of the individual source images. Initially fusion of MRI and CT scan images has been carried out using Stationary Wavelet Transform (SWT). Then watershed transform is applied for image segmentation and discriminative robust local binary patter (DRLBP)is employed to extract the features exactly from the fused image. Classification of the tumor is done by Support Vector Machine (SVM) thereby reducing the generalization error and increasing the accuracy. The ultimate goal is to classify the tissues into normal and abnormal using machine learning algorithms .Image fusion process yields more accurate information of the brain than any of the individual source images.