Early brain tumour prediction using an enhancement feature extraction technique and deep neural networks

Abstract

Early tumor detection in the brain plays a vital role in early tumor diagnosis and radiotherapy planning. Magnetic resonance imaging (MRI) is latest technique which normally used for assessment of the brain tumor in Hospitals or scan centers. MRI images are used as the input image for brain tumor detection and classification. For predicting brain tumor earlier, an enhancement feature extraction technique and deep neural network are proposed. At first, the MRI image is pre-processed, segmented and feature extracted using image processing techniques. Support Vector Machine (SVM) based brain tumor classifications were performed previously with less accuracy rate. By using DNN classifier, there will be an improvement in accuracy rate. The proposed method mainly focuses on six features that are entropy, mean, correlation, contrast, energy and homogeneity. The performance metrics accuracy, sensitivity, and specificity are calculated to show that the proposed method is better compared to existing methods. The proposed technique is used to detect the location and the size of a tumor in the brain through MRI image by using MATLAB.