LAUSR

With the advent of image processing technologies, the in-depth portion of human body can be epitomized visually to perceive abnormalities in human anatomy. Image processing is a tool for identifying the substances and obtaining information from them. Medical image processing is a stimulating area to diagnose diseases specifically, brain cancer, breast cancer, liver cancer, neuro- and cardio-diseases, etc. Image segmentation is an act of segregating the images into various parts to identify a particular substance and its margins. Brain tumor is the irregular and intense growth of tissues causing cancer. The most used technique to diagnose brain tumor is Magnetic Resonance Imaging (MRI). Precise information about the affected area is crucial for the appropriate treatment. As numerous data are created in MRI diagnosis, an automated segmentation technique is necessary to obtain precise information of tumor. In this paper we presented Depth-First Search (DFS) segmentation algorithm based on graph theory. Here the image pixels are arranged into a tree like structure based on their proximity in the image. The experimental results are compared with other existing systems. Also performance measures of ANFIS classifier and SVM classifier are compared. It distinguishes healthy cells from the cells affected by brain tumors. In the proposed method, the computational complexity is reduced and accuracy is enhanced.