Brain Image Classification Using Dual Tree M-Band Wavelet Transforms and Support Vector Machine

Human brain is an important organ of humans, and it is the center of the nervous system. Brain tumor causes more damage to humans, and the cause is still unknown. Hence, an early diagnosis system for a brain tumor is required. In this paper, an efficient method is presented for brain image classification using Dual Tree M-band Wavelet Transform (DTMWT). The system uses Magnetic Resonance Imaging (MRI) images of brain. REpository of Molecular BRAin Neoplasia DaTa (REMBRANDT) database is used for MRI brain image classification. DTMWT uses the Hilbert pair of wavelet for decomposition. A predefined number of DTMWT sub-band coefficients are selected and directly fed to the Support Vector Machine (SVM) classifier. Results show that DTMWT and SVM classifier provides 86% accuracy, 79% sensitivity and 93% specificity for MRI brain image classification system.