Enhanced Deep Learning Model Based Brain Tumor Segmentation Using M-SVM

An uncontrollable increase in abnormal cells in brain tissue produces the brain tumors. There are two types of tumors. They are benign and malignant. Adjacent cells are not affected by benign type of tumor cells, whereas malignant type tumor affects the adjacent cells. Tumors creates serious life-threat to humans. So, it is necessary to identify this tumor in initial stages and get treated. In order to identify and segment brain tumors, an enhanced deep learning model based brain tumor segmentation using SVM is proposed in this work. Magnetic Resonance Image (MRI) images are used in the experimentation. Captured MRI images are pro-processed using Laplacian of Gaussian Filter (LoG). Features are extracted from pre-processed image. Based on the extracted features MRI images are classified as normal or abnormal. If an image is classified as identified are abnormal, deep learning method is used to segment the tumor regions. Better results are obtained by the proposed technique as demonstrated in experimental results.