MRI images are the most noninvasive technique for tumor detection. 3D MRI images are difficult to process because of their size. 3D MRI images have three orthogonal planes (TOPs): axial plane, coronal plane, and sagittal plane. Each plane carries information about the tumor. In this research work, brain tumor is detected and classified as malignant and benign from the features extracted from these orthogonal planes. Local Binary Pattern (LBP) is used for feature extraction. All modalities of MRI images, i.e., T1, T2, T1ce, and Flair, are used for feature extraction. The experiments are performed on the Brats 2018 dataset with SVM and KNN classifiers. In one experiment, features retrieved are given to train the SVM classifier. In the second experiment, the retrieved features are given to train the KNN classifier. By using different combinations of malignant and benign MRI images with SVM and KNN classifiers, it is found that SVM performs better than KNN classifier.