Brain Tumor Segmentation Using Deep Convolution Neural Network and Support Vector Machine

Abstract

Tumor segmentation is the primary and tedious task for the clinical experts. Computer Aided Design is the only solution which identifies the tumor very accurately with less time. Deep learning models such as the convolutionary neural network have been widely used in 3D biomedical segmentation and have achieved state-of-the-art performance.In this research, saliency based deep features are extracted from MRI. Then Support Vector Machine is used for classifying deep features. The proposed method is tested on BRATS 2015 dataset and it is compared with state-of-methods and recent methods. The proposed method achieves 0.94, 0.93 and 0.9 as dice score, precision and sensitivity respectively which is greater than other methods.