Comparative Analysis of Segmentation Algorithms for Brain Tumor Detection in MR Images

Tumors can occur in any parts of the body. A brain tumor can be considered as one of the genuine and hazardous tumors. It is really made either by the strange and uncontrolled cell division inside the brain or from diseases essentially display in different parts of the body. By and large, tumors are characterized in light of the area of their cause and its threat. We proposed to set up a no holds barred examination of three brain tumor division procedures. We are displaying correlation of district developing, split and union and K-Means cerebrum tumor division methods. We have performed all our analysis and simulations using MATLAB R2013B. We have chosen medical dataset BRATS2016_Training for our thesis analysis. The dataset is specially based on brain tumor medical images. We have selected images for our research from this data set and performed our research. From our research work we concluded that region growing stands to be best technique for brain tumor segmentation on the basis of parameters derived. The parameters derived are Area, Region of Interest mean color (ROImean), Rest mean color (Restmean), Region of Interest Standard Deviation (ROIstD), Rest Standard Deviation (ReststD), Region of Interest Peak to Mean Ratio (ROIPMR), Rest Peak to Mean Ratio (RestPMR), Convex Area (ConArea), and Aspect Ratio (AsRatio)