Research On Segmentation Algorithm Of Triangle Mesh Model Of Dental Crown

With the development of computer-related technologies,oral materials and the ma-turity of 3D printing technology,Computer Aided Design and Computer Aided Manu-facturing（CAD/CAM）systems have gradually been introduced into the dental medical field.Dental medical software algorithms have gradually become the focus and focus of research,and crown segmentation is the basis and focus of dental medical CAD/CAM systems.In this thesis,the segmentation algorithm of the triangular mesh model of the dental crown is studied,and the main work is as follows:The curvature analysis was performed on the crown triangular mesh data,and it was concluded that the curvature of the transition regions between crowns of different teeth and the midpoints of the transition regions between crowns and gingiva were smaller.Therefore,all three algorithms in this thesis have used this feature to complete the crown segmentation.A crown segmentation algorithm that combines mesh subdivision and region growth is proposed.The algorithm first uses Loop subdivision to smooth the mesh and smooth the segmentation path,and then manually interacts with the initial seed point for each tooth,using two types of curvature The degree evaluation functions H₁and H₂replace the accurate calculation of the curvature,and the region growing algorithm grows from the seed point to the boundary of the crown.Among the 14 models,the segmentation accuracy of the two curvature evaluation functions is not much different,both above 96%,but the curvature evaluation function H₂has fewer manual interactions than H₁.Improved a crown segmentation algorithm based on the shortest path.After the Loop subdivides the mesh,the algorithm defines new calculation formulas for curvature and edge weights,and obtains multiple marked points through manual interaction,and then uses the shortest path algorithm calculate the shortest path between the marked points,and the closed loop formed by the shortest path between all the marked points is the par-tition path of the crown.Among the 14 models,the average segmentation accuracy of the improved algorithm is 99%,which can better segment teeth with more depressions.A new morphological crown segmentation algorithm is proposed.After the Loop subdivides the mesh,the algorithm first proposes an automatic curvature threshold method to extract the transition area,and then increases the connectivity of the transition area through morphological expansion,and then the transition area use principal component analysis（PCA）to fit the plane,remove the transition area below the plane,and then use manual interaction to ensure that the transition area is completely closed and disconnect some areas,and then use the region growing algorithm to remove small transition areas,and finally use the morphological erosion iterative algorithm to extract the skeleton line from the transition area,which is the segmentation path of the crown.Among the 14 mod-els,the average accuracy of this method is 95.91%.Compared with the region growing algorithm,this method has a smoother segmentation path.Compared with the shortest path algorithm,this method has more convenient manual interaction and fewer interac-tions.Finally,a simple crown segmentation software is designed and implemented,which integrates the three algorithms of this thesis.The crown segmentation system can accu-rately and conveniently segment the crown and help users improve their work efficiency.