A Full Crown Restoration Approach for Defect Teeth Based on CGAN

The traditional manual restoration for defect teeth is time-consuming and the crown restoration made by dental CAD/CAM systems lacks personalized anatomical features.This paper proposes a novel method for full crown restoration based on conditional generative adversarial networks and high-dimensional features loss function.First,the 2D depth maps of the tooth preparation and conditional models are obtained by calculating the depth information of the 3D model,and then the dataset for the network model is constructed.Second,the network model applies the opposing tooth to constraint occlusion relationship and uses the symmetrical tooth with same position as supplementary information,and then the network model maps the tooth preparation to the crown designed by technicians.After that,the tooth crown depth map generated by the network model is reconstructed into a three-dimensional mesh model.Finally,some dental models are selected for the experiment,the effects of different constraints on the occlusal surface morphology of generated crowns are analyzed and the quality of crowns repaired by different methods is compared.The experimental results show that this restoration method can reconstruct the personalized anatomical features on the crown occlusal surface efficiently and fulfill the requirements of full functionality restoration principles for the defect tooth.