A Brain Tumor: Localization Using Bounding Box and Classification Using SVM

The brain tumor is defined as the abnormal growth of unhealthy and unnecessary cells in the brain. The objective of the proposed method is to identify and locate the presence of tumor in the Magnetic Resonance Imaging (MRI) of brain images. The proposed method incorporates three phases to determine the presence of brain tumor, namely, preprocessing, identifying/locating the tumor region, and classifying the tumor region. The input image is filtered to reduce the noise in the preprocessing phase. In the second phase, Bounding Box (BB) is used to identify/locate the tumor region in the filtered image. Subsequently, in the third phase, Support Vector Machine (SVM) is used to classify the exact tumor location. Finally, the brain tumor is localized absolutely by the proposed tumor detection method. Moreover, the proposed method is evaluated with the publicly available standard dataset and compared with a contemporary method. The experimental results concluded that the proposed method has higher tumor detection accuracy than the existing method.