An Automatic Classification of Magnetic Resonance Brain Images Using Machine Learning Techniques

摘要

In medical image analysis, an automatic Magnetic Resonance (MR) brain image classification is very important. This research paper's objective is to classify the abnormality of MR brain image. This paper presents a Discrete Wavelet Packet to separate highlights from images, followed by using principal component analysis to decrease the measurement of highlights which were submitted to a Kernel support vector machine(KSVM). The procedure of K fold stratified cross-validation was utilized to upgrade the speculation of KVSM. The proposed method has been tested by 160 Magnetic Resonance brain images, both normal and abnormal which are collected from the Harvard Medical School website; it achieves better accuracy of classification as 99.38%. The performance of this proposed method has been evaluated with four different kernels. The comparison has been done by various states of the art methods. From the experimental data, our method was effective and rapid. This classification method can help doctors to diagnose the patients.