ABSTRACT


The segmentation of brain tum or using Magnetic Resonance Image (MRI) plays an important role m the medical image process. The mam reason behind this process is to

separate the different types

0f t'
issues such as necrotic core, active cells and edema

from normal brain tissue of h ·t
w 1 e matter (WM), gray matter (GM) and cerebrospinal
fluid (CSF). Presently the b f . . . .
num er o patients of brain tumor are increasing
significantly and the treatment of brain tumor is highly expensive. The rad10log1st an
neurologist have to do a tough job for the detection and treatment of brain tumor in the patients. Ultimately they need fast and efficient treatment methods. A number of techniques to detect the brain tumor have been proposed in literature each having merits and demerits. In this study a new automated system is proposed for brain tumor detection in which features of MRI images are extracted by using segmentation and morphological operation. The proposed framework consists of three phases; in the first phase the input MRI image is preprocessed to remove the noise and sharpen the image for the purpose of better segmentation of the tumor. In the second phase, the threshold segmentation is applied to extract the tumor region and quantify the tumor area in terms of square per inch and the number of pixels reserved by the tumor regions present in the brain. Finally, in the third phase the post-processing is done to remove the false regions so that the extracted tumor is better understandable to the radiologist. It was revealed that the proposed technique offers tumor region by consuming less time in milliseconds and low computational cost and classifies the tumor region in the forefront for the considerable visual investigation that has a percentage difference of
le than 1%.


Keywords: Magnetic Resonance operation.

Image

(MRI)•

Segmentation, Morphological