Robust Classification of Brain Tumor in MRI Images using Salient Structure Descriptor and RBF Kernel-SVM

Brain tumor classification in Magnetic Resonance Imaging (MRI) is vital as it delivers data about unusual tissues which is essential for planning treatment. Automating this method is challenging due to the high variety in appearance of tumor tissue among dissimilar patients and in many circumstances, comparison between tumor and normal tissue. In this paper, we develop a brain tumor classification algorithm based on mutimodel texture features and computational intelligent technique. The proposed method consists of three stages. In the first stage, the brain MRI is segmented into white matter (WM), gray matter (GM), Cerebrospinal fluid (CSF) and outliers by ordering of observations. In second stage, various texture and intensity histogram based features are extracted using salient structure histogram. In the third stage, the extracted features are trained using pair of RBF kernel SVM. The classification accuracy of the proposed HSDH+ HKSVM is 96.8%, but the existing methods HSDH+RBF and HSDH+SVM produce 91.32% and 94.32% respectively. The developed methods for feature extraction, and classification of MRI image can be amalgamated to develop a CAD system. This system would be beneficial to radiologists for precise localization, diagnosis, and interpretation of lesion types of diabetic retinopathy in retinal images.