Degenerative disc disease is an age-associated disorder which leads to chronic back pain. This paper proposes computer aided diagnosis on intervertebral disc degeneration and its classification. This work aims to classify intervertebral disc (IVD) of spine Magnetic Resonance Imaging (MRI) as normal (healthy) or degenerate disc by considering the intensity, texture and shape features. Phases of the work include pre-processing, segmentation, feature extraction. Proposed algorithm uses features of gray levels, texture, and shape of IVDs, and further to classify IVDs as normal or degenerate. Several classifiers were used to detect disc as healthy or degenerate. Classifiers adopted in the proposed work includes: Support Vector Machine (SVM), Linear Discriminant Analysis (LDA) and Naive Bayes classifiers. SVM is preferred for the classification of disc as normal or degenerate with the classification accuracy as 81.81%. Performance evaluation of this work uses 15 Spine MRI. Experiments on real spine MRI that has been taken from publicly available Spine database show the effectiveness of our approach.
Volume 11 | 07-Special Issue
Pages: 1075-1087