A novel approach for the detection of tumor in MR images of the brain and its classification via independent component analysis and kernel support vector machine

Introduction Brain tumor is due to the growth of abnormal cells in the brain. Brain tumor in its final stage is converted as brain cancer, which leads to death. Survival rate can be increased if the tumor is detected and diagnosed in the initial stages. The detection of the tumor in MR images of the brain can be done using segmentation. These brain tumors exist in different types which make the diagnosis decisions very difficult. Hence to provide best and right treatment for the tumor it is very important to classify the type of tumor. Brain tumor can be benign or malignant. Benign being noncancerous and is treated as a low-grade tumor. Malignant is cancerous and treated as a high-grade tumor. A Benign tumor is less harmful than malignant.