Graph theory based image segmentation for Brain Tumor detection

Abstract

In recent years, brain cancer is a significant reason for the mortality of a larger section of newlinethe population across the globe. Performing the segmentation of tumors manually by newlineradiologists or physicians is a tedious job and requires a lot of time because of the huge amount newlineof clinical data that is produced today in hospitals. Among the imaging modalities, MRI plays newlinea major role in the visualization of the anatomical structure of the brain. The main contribution newlineto this proposed work is the development of segmentation and classification algorithms. newline newlineThis research work performs pre-processing of input images, to improve the quality of newlineimage. Pre-processing techniques like region of interest, inverse methodology, boundary newlinededuction method, and median filter is used to remove noises and unwanted data in an image. newlineRegion of interest identification includes distinguishing an image region that seems totally newlinedifferent from the rear ground with regard to options like distinction, color, region, size, form, newlineand texture pattern. The inverse technique supports spotting the lower and higher boundary of newlinethe digitally regenerated values. In the boundary technique, the outer space of the image is newlineremoved, and solely the region of the known image is taken for clustering. The median filter newlinecan remove the noise as well as sharp edges present in the MR image by preserving the edges newlinecharacters. The pre-processed image using the mentioned techniques will be used in the newlinesegmentation and classification process. newline newlineThe first contribution is the development of a graph theoretical approach for the newlinesegmentation of brain tumors. In this method, graph theory and dynamic programming methods newlineare combined to develop the segmentation algorithm. Graph theoretical approaches are newlineextensively used to detect boundaries of tumors in MRI images of the human brain due to their newlineflexibility in representing any type of image and minimization representation of images. In newlinegraph theoretic methods.