Automatic Tooth Region Separation for Dental CT Images

In this paper we focus on the technique to construct and visualize the individual tooth model from CT image sequences for dental diagnosis and treatment. For CT images the teeth are touched with each other in some slices. We have to find the individual region for each tooth and separate two teeth if they are touched together. We propose method based on the distinguish feature of oral cavity structure. We utilize the full 3D data instead of 2D projections which may lose some information. We firstly separate upper and lower tooth regions and then fit the dental arch using fourth order polynomial curves after a series of morphological operations. We assume that there exists a plane separating two adjacent teeth in 3D space. In this plane the integral intensity value reaches minimum. Along each arch point we obtain a plane and calculate the integral intensity. We use these values to draw a profile and by analyzing all the local minima we get the separating point and the position of the separating plane. The information for the tooth region can guide us to segment the individual both tooth contours in 2D and tooth surface in 3D space. Experiments for both separation and visualization of each tooth verify our method is automatic and robust.