Classification of Brain Tumor on CT Images using Texture Analysis

Abstract

Diagnosis of brain tumor is very complicated and critical process usually due to the precautions of the invasive operation on this area. This is analytical study aimed to classify the brain tumors using images textural analysis the result of this study presents a highly efficient method to identify, diagnosis and classify the intracranial tumors. The study had been worked up on one hundred and nine patients have different types of brain tumors, it done in Khartoum and Gazera state using different types of CT machines including GE and Neusoft scanner. This study consists of two methodology; first order statistical analysis and higher order statistical analysis for images classification. For the first order statistical analysis the procedure contain dual phase, phase one contain textural feature extraction from the selected images (features included: mean, skewness, energy and entropy), phase two classification of images component according to their extracted textural features using linear discriminant analysis, the results showed that the proposed method have achieved high accuracy in recognizing the intracranial tumor from the normal surrounding tissues. The accuracy of classification was 90.4%. For the higher order statistical analysis nine features include; long run emphasis, low gray-level run emphasis , high gray-Level, Run emphasis, short run low gray-level emphasis , short run high gray-Level emphasis, long run low gray-level emphasis, long run high gray-level emphasis, gray-level non-uniformity, run length non-uniformity, run percentage had been used for image classification. The analyzed data showed accuracy percentage beyond to 94%.That exploring the priority of higher order statistical analysis method upon the first order on classification of brain tumor when the classification process based on the tumor textural feature.