3D reconstruction for maxillary anterior tooth crown based on shape and pose estimation networks

Area of Science:

• Digital dentistry
• Computer-aided design (CAD)
• Biomedical engineering

Background:

• Maxillary anterior tooth crown design is critical for aesthetics.
• Current manual or semi-automatic methods are time-consuming.
• Existing automatic methods lack relational crown analysis and focus on occlusal surfaces.

Purpose of the Study:

• To propose a novel method for automatic 3D reconstruction of maxillary anterior tooth crowns.
• To improve the efficiency and accuracy of digital dental crown design.

Main Methods:

• Development of a pose estimation network (PEN) and a shape estimation network (SEN) for joint point cloud estimation.
• SEN utilizes a transformer encoder to capture inter-crown shape relationships for precise reconstruction.
• Iterative refinement of point clouds, incorporating pose transformation, to generate the final crown mesh model.

Main Results:

• The method was evaluated on 600 cases, with both networks converging within 1000 epochs.
• Achieved an average deviation of 0.22 mm for point cloud reconstruction and 0.13 mm for the final mesh model.
• Reconstruction time per case is under 11 seconds.

Conclusions:

• The proposed method enables automatic and accurate 3D reconstruction of missing maxillary anterior tooth crowns.
• Demonstrates significant improvements in work efficiency compared to traditional methods.
• Shows promising application prospects in digital dentistry and restorative treatments.