ZenToks Publication

Chapter 6

A Review of MRI Based Automatic Brain Tumor Detection and Segmentation

Abstract/Preface

In medical image processing techniques, detection of brain tumors from the MRI images plays a very challenging role. It is considered a very powerful imaging technique which is capable to diagnosis the abnormalities in the brain compared to other medical imaging techniques such as X-ray, Computed Tomography (CT), Positron Emission Tomography (PET) etc. Experimental studies on MRI based brain tumor segmentation are getting more attention and coming closer to clinical acceptance as it provides non-invasive images with high resolution and excellent contrast between the different soft tissues of the body. Quality of brain images are affected by several problems like noise and partial volume effect due to overlapping tissues. These problems need to be addressed for accurate segmentation, which is extremely important and essential for exact diagnosis by computer aided clinical tools. Different methods have been developed for segmentation of brain tumor efficiently. The most important application of segmentation technique is to isolate the tissue of the tumor part which includes active cells, necrotic core and edema from the normal part of brain tissues consisting of White Matter (WM), Gray Matter (GM), and Cerebrospinal Fluid (CSF). The purpose of this paper is to provide a comprehensive review of different brain tumor segmentation methods using MRI images by studying their advantages and disadvantages with earlier proposed segmentation techniques. Firstly, a brief introduction about brain tumors, their types and the reasons for brain tumors have been introduced. Then the comparison of different imaging modalities has been presented. Lastly a final assessment has been made by addressing the future developments and trends for MRI-based brain tumor segmentation methods.