Segmentation and Isolation of Brain Tumors Using Different Images Segmentation Methods

Abstract

Brain tumors are an anomaly growth or mass of cells in or around the brain tissues. Brain tumors are two types either malignant (cancerous) or benign (noncancerous), all of these can affect children and adults. Tumors of the brain can impact human brain function if they grow large enough to press on surrounding tissues. Brain tumor is an inherently serious and life threatening. The brain tumor diagnosis is depending on the specialist, but it may give different diagnoses which may vary from one specialist to another depending on the accurate diagnosis of tumor size. To avoid that, in many cases computer is used as an aided method for the segmentation of brain tumor. Segmentation in image processing is the process of dividing regions into different parts depending on various criteria such as intensity, homogeny and other.  In this study, many segmentation techniques are used to segment the brain tumors of MRI images: Edge detection methods (Sobel, Prewitt, and Canny); thresholding methods using different threshold (T) values; and K-Means clustering methods with different numbers of cluster (K). From the results of segmentation methods, it is clear that, edge detection methods are failed to segment and isolate the region of tumor, thresholding segmentation method gives a good result when using T= 150,200 for some cases. While K-Means segmentation method is successful to segment and isolate the tumor of the brain for one case when using two clusters (K=2).

Received 02/09/2022

Revised 07/05/2023

Accepted 09/05/2023

Published Online First 20/01/2024