MRI Based Brain Tumor Detection Comprising a Hybrid Segmentation Technique

The growth of abnormal tissue in the brain causes brain tumors. There are two main types of tumors, such as-non-cancerous (benign) and cancerous (malignant). Because of the several unique types of brain tumors, it is difficult to forecast the survival of a patient. Tumor classification and detection is a vital sector in medical science and diagnosis of patients. The usual technique of detecting brain tumors is a Magnetic Resonance Imaging (MRI) scan. However, it is an iterative and human interactive task that greatly varies depending on the personal expertise of radiologists and clinical experts. Simple image processing method sometime fails to expose the problem of the human organ. In this issue, researchers are trying to adopt the emerging techniques of Machine Learning, and they overcome many challenging issues by using Machine Learning (ML) algorithms. This work implies the detection and classification of MRI brain tumors by feature extraction and ML classifiers. After prepeocessing and feature extract of MRI image classifiers are applied. Laplacian filter, Discrete Wavelet Transformation, and Principal Component Analysis are used for the preprocessing stage. Finally, kernel support vector machine (KSVM) is used to effectively detect and classify brain tumors as benign or malignant with an accuracy of 89%. This research has thrown light on the field of diagnosing brain tumors.