Certain Investigation on Brain Tumor Segmentation and Classification On Mr Images

Abstract

ABSTRACT newlineMedical imaging has become as a transparent discipline in diversified medical diagnosis. It plays a vital role in automatic detection, which bestows information about abnormalities for further treatment. The traditional approach of detecting MRI has been based on manual inspection, which has become inappropriate for vast volume of data. Automated tumor detection has been gaining importance that conserves the time of the radiologist. In this research work, the pre-processing of MR Images is carried out by Wiener filter and the segmentation has been done by utilizing Adaptive Connected Component Pixel Segmentation technique. The Support Vector Machine (SVM) is proposed for MR brain image classification to identify whether the image is normal or abnormal. The performance of SVM is compared with those of neural classifiers such as Back Propagation Neural Network (BPNN), Levenberg Marquardt (L-M) and Gradient Descent (GD)- based BPNN algorithms. The result reveals that the proposed method produces better results in terms of accuracy for MR brain tumor classification. newlineBrain tumor is a life threatening disease. The brain contains more than 10 billion working brain cells. The damaged brain cells are diagnosed themselves by splitting to make more cells. This regeneration takes place in an orderly and controlled manner. If the regeneration of the cells gets out of control, the cells will continue to divide developing a lump which is called tumor. In proposed method feature extraction is done by hybrid method. In this paper support vector machine and naïve bays is used for detection of tumor and non tumor Image segmentation by using fuzzy c-means Segmentation Method. In this paper the Gray Level Co-occurrence Matrix (GLCM) is used as a feature. Main advantage of this method is that it will give fast and accurate result with the help of training data set and it reduces time and computation power. newline