Segmentation of Premolar Based on Geodesic Active Region

To find borders between homogeneous regions in the various morphologies, several segmentation resulting from edge-based and region-growing can not produce exactly the same, and a combination of results often work more accurately. Also, clinically usable segmentation for medical imaging requires a high degree of interaction with registration algorithms such as the Insight Toolkit (ITK) and the Visualization Toolkit (VTK). In this paper, we proposed the geodesic active region segmentation to find the inner structure of premolar teeth acquired by microcomputed tomography (micro-CT) scanner. As a result, we discriminated enamel, dentin and pulp zones. Furthermore, we showed that the 3D geometric models of premolar would be useful for the tooth morphology.