Automatic Diagnosis of Disc Herniation in Two-Dimensional MR Images with Combination of Distinct Features Using Machine Learning Methods

Low Back Pain (LBP) is one of the most common reason why people see a spine specialist. This back pain is often caused by lumbar disc herniation. Computer Aided Diagnosis (CAD) systems help reducing medical errors and improving health care quality, like speed and accuracy in diagnosis by radiologists. In this study, we used 50 clinical Magnetic Resonance Images (MRI) cases including 250 lumbar area discs. A diagnosis system have been developed to locate, label and segment discs by processing T1-weighted and T2-weighted sagittal view of MR images in order to diagnose herniation. Then, we extracted some distinct features such as shape, intensity, and geometric features of the discs. In the next step, we used k-Means clustering and active contour model (snake) on the region of interest (ROI) for segmenting discs and extracting features. We find the combination of effective features. The result demonstrated an average of 97.91% and 97.08% accuracy with K-fold cross validation method using k-Nearest Neighbor (KNN) and linear Support Vector Machine (SVM) classifiers respectively, to show a robust method.