Feasibility study of machine vision for diagnosis of multiple sclerosis

Abstract

Multiple sclerosis (MS) is a chronic autoimmune disease which affects various areas of inflammatory demyelination within the central nervous system. The symptoms of MS are muscle weakness, weak reflexes, muscle spasm, difficult to move, miss-coordination and unbalance with others. Image processing plays a vital role in diagnosing multiple sclerosis. This paper proposes a novel method to detect the brain disease multiple sclerosis by segmenting brain into right and left hemispheres which are then divided into 16 quadrants. The disease is identified by the presence of higher pixels in specific quadrants of white matter and regions near to Cerebro spinal fluid(CSF); also by the absence of higher intensity pixels on gray matter. A total of 10 multiple sclerosis and healthy brain images are taken for experimentation. The result obtained show that the proposed method successfully identifies multiple sclerosis.