ABSTRACT
A collection of abnormal cells within the brain is called as brain tumor. The tumors can be cancerous or non-cancerous. A neurologist or a neurosurgeon may use computer tomography (CT) scan or magnetic resonance imaging (MRI) for diagnosis of a brain tumor. The test results are then used by the neurologists to prescribe treatment for their patients. Segmentation of medical images for detecting the exact size or location of tumor is challenging due to the intrinsic nature of images. There are a number of segmentation methods that have been developed to overcome the difficulties face by the radiologists and the doctors to detect the abnormalities in the medical images due to noise or due to spatial variations in illumination. Practical swarm optimization (PSO) techniques are being used by the researchers to lessen the effects of challenges faced while detection of brain tumor from MRI images and reduce the time required to give exact results of the tests. PSO when combined with other soft computing techniques like Fuzzy and genetic algorithm can improve the quality of segmentation for the MRI images. Though beneficial for the people who need them but it may require long time one hour to complete the whole process. Machine Learning (ML) can reduce the time needed for producing MRI images.
