Soft computing based classification algorithms for MRI brain images using rough set theory and texture features

Abstract

In general the frequently used medical imaging method is Magnetic Resonance Imaging (MRI). Various methods have been stated for diagnosing tumor in MRI brain images, most particularly, feature extraction and feature classification algorithms. Many feature extraction techniques are available to extract the inter and intra tumor features. Similarly many feature classification algorithms are available for differentiating tumor of various types. In this research work the several combinations of feature extraction and classification algorithms were used to analyze the best pair of extraction and classification algorithms. The key objective of this research is to propose novel brain tumor identification system based on the pair of extraction and classification algorithms. Feature extraction consists of both micro and macro-scale texture features encountered in MRI images. The proposed feature classification algorithms include kernel and optimization techniques to increase the efficiency of the brain tumor identification task. newlineThis research work concentrates on developing a Computer Aided Detection (CAD) System using MATLAB software. It is a tough task to diagnose brain tumor using MRI images. For easiest diagnosis of brain tumor in MRI image, an automated system is designed which decreases the number of false readings, both positive and negative. The automated system enhances the chance of diagnosing abnormalities at the earliest. Recently soft computing techniques play a very important function in practical use of medical field. It differentiates the normal and abnormal MRI brain images and helps in earliest diagnosis of brain tumor. newline newline