Full Tooth Contour Recognition and Model Reconstruction Method Based on CT Images

It is of great significance for the treatment of oral diseases to establish a digital 3d model of teeth based on oral CT images. In order to establish a 3d digital model of teeth by using oral CT images, the contour segmentation, coordinate transformation and surface model reconstruction of oral CT images are required. The structure tensor information can reflect the local structure information of the image by using the multi-channel information, and many models constructed by using the local tensor information have the de-noising effect. Therefore, the area term fused with structural tensor information was added to the (Local Gaussian Distribution Fitting) LGDF model to segment the contour of a single tooth. The segmentation results show that the improved model can be segmented into ideal contours without Interference noise area and false boundary. After the segmentation of all the contours was completed, all the contours were converted from two-dimensional CT coordinates to three-dimensional coordinate forms, and the surface model was obtained by surface reconstruction of three-dimensional scattered point cloud using Crust algorithm. The surface model construction method proposed in this paper has better surface quality.