Detection of Brain Tumor Using Image Processing

Abstract

Brain tumor is an accumulation of anomalous tissue in the brain. Tumors are primarily classified into malignant and benign when they develop. It can be life threatening hence it is important to recognize and identify the presence of tumors in brain image. This paper proposes a system to decide whether the brain has tumor or is it tumor-free from the MR image using combined technique of K-Means and support vector machine. In the first stage the input image is converted to grey scale using binary thresholding and the spots are detected. The recognized spots are represented in terms of their intensities to distinguish between the normal and tumor brain. The set of feature extracted are later characterized by using K-Means algorithm, then the tumor recognition is done using support vector machine.