Brain Tumors Classification by Using Gray Level Co-occurrence Matrix, Genetic Algorithm and Probabilistic Neural Network

Abstract

Background:Brain tumors classification by MRI (Magnetic Resonance Imaging) is important in medical diagnosis because it provides information associated with anatomical structures as well as potential abnormal tissues necessary for treatment planning and patient's case follow-up. There are a number of techniques for medical image classification. In this paper brain tumors detection and classification system are developed into seven tumors types. The image processing techniques such as preprocessing by using a mean filter and feature extraction have been implemented for the detection of a brain tumor in the MRI images. In this paper, extraction of texture features using GLCM (Gray Level Co-occurrence Matrix). We used Probabilistic Neural Network Algorithm (PNNA) for image classification technique based on Genetic Algorithm (GA) and K-Nearest Neighbor (K-NN) classifier for feature selection is proposed in this paper.

Objective: MRI brain tumors detection and classification system by using GA and PNN which able to diagnose different types of tumors in human brain.

Patients and Methods: Medical image techniques are used to imaging the internal structures of the human body for medical diagnosis. Image processing is an effective field of research in the medical field. MRI dataset, obtained from the Atlas Website of Harvard University.

Results: Brain Tumors are classified by using the genetic algorithm where the total number of features (20 features) has been reduced to 10 features as the strongest features in the classification.

Conclusion: MRI brain image is one of the best methods in brain tumor detection and classification, by observing only MRI images the specialists are unable to keep up with diagnosing. Hence, the computer-based diagnosis is necessary for the correct brain tumor classification.