Brain Tumor Detection and Classification of MRI Brain Images Using Morphological Operations

Purpose Image processing is a vital aspect of medical science which enables visualization of various anatomical structures of human body. Planar imaging can be used for detecting and visualizing hidden abnormal structures which are not use to visualize using simple imaging. Magnetic resonance imaging (MRI) modality is one of the techniques which enables scan and capture of internal body soft tissues. This work describes the process implemented for detection and extraction of brain tumor from patient’s MRI scan images of the brain. Procedure The process includes some contrast enhancement, noise removal functions, segmentation, and morphological operations which are the basic terms of image processing. By using MATLAB software, we detected and extracted tumor from 24 MRI scan images of the brain. We calculated the tumor properties including area, perimeter, and eccentricity. Using those properties, we then used k-medoid clustering for classification. Results Detection of tumor was performed on 24 MRI brain images, and their properties were calculated. The images that have maximum similarity and show the characteristics of a benign tumor type and few of the tumors have malignant characteristics. Conclusion The work finds significant application in diagnosis of epilepsy, cancer, radiotherapy, etc.