Automatic Segmentation of Intervertebral Discs of Cervical Spine Magnetic Resonance Images

Abstract

ABSTRACT newlineThe cervical spine is more mobile than the other spinal regions. It contains Inter-Vertebral Discs (IVDs), vertebrae and spinal cord. The IVDs are a gel-like cushion that lie between each vertebra and prevents the vertebrae from rubbing against each other. It provides the shock absorbing function and protects the spine during the normal activities like walking and running. The cervical disc diseases are normally diagnosed by radiologists or the doctors by considering the cervical MR images. The master slice of cervical MR images contains all the organs like skull, vertebra, trachea and spinal cord. Accurate and earlier diagnosis of all the disc abnormalities is only possible if the disc is completely segmented from the neigh boring organs. Clinical diagnosis currently relies on visual interpretation of radiologists and doctors. In clinical practice, the evaluation of disc degeneration is based on qualitative descriptions of disc features such as the signal intensity and height of the disc s nucleus. This qualitative evaluation suffers from inter-and intra-observer variability. Also, earlier diagnosis of disc abnormalities are very difficult in both semi-automatic and manual segmentations of IVDs. Therefore, automatic segmentation is the best solution to overcome all the above-mentioned problems. The automatic segmentation of IVDs is highly challenging due to the intensity resemblance between discs, ambiguous disc boundaries, different sizes and shapes of discs, vertebra and spinal cord. newlineResearchers are working on the Computer Aided Diagnosis (CAD) of discs abnormalities. The CAD system for the diagnosis of discs abnormalities requires segmented discs for the analysis. But most of the previous works had concentrated only on the labelling of the discs and only a few attempts has been made to segment the discs automatically from the other anatomical structures. This research work is focused on developing new automatic segmentation algorithm for segmenting the IVDs of cervical spine MR images. This work prop