<![CDATA[Brain tumors are a global health problem, ranking 12th as a cause of death. MRI is used in the diagnosis of brain tumors because of its ability to display soft tissue structures in detail, but manual interpretation of MRI images by radiologists is still subjective. Therefore, a more objective computer-based classification approach is needed. One factor that could potentially affect classification performance is the difference in MRI image orientation, namely axial, sagittal, and coronal. This study aims to analyze the effect of MRI image orientation on GLCM and SVM-based brain tumor classification. The preprocessing stage includes cropping, noise reduction, and resizing. Feature extraction was performed using GLCM with distance d = 1 at angles of 0°, 45°, 90°, and 135° with contrast, correlation, energy, and homogeneity features. Classification was performed using SVM with Linear, Polynomial, RBF, and Sigmoid kernels. The test results show that the axial orientation produces the highest accuracy of 78% with the Linear kernel, the sagittal orientation achieves an accuracy of 83% with the Polynomial kernel, and the coronal orientation provides the highest accuracy of 86% with the RBF kernel. These findings indicate that the orientation of MRI images affects the performance of texture-based brain tumor classification.]]>
Copyrights © 2026
