DTR-Net: Dual-Space 3D Tooth Model Reconstruction From Panoramic X-Ray Images

Area of Science:

• Digital dentistry
• Medical imaging
• Computer vision

Background:

• Cone-beam computed tomography (CBCT) provides 3D tooth models but involves high radiation doses and costs.
• Reconstructing 3D tooth models from 2D panoramic X-rays is a cost-effective alternative.
• There is significant clinical interest in developing efficient 3D tooth reconstruction methods.

Purpose of the Study:

• To propose a novel dual-space framework, DTR-Net, for reconstructing 3D tooth models from 2D panoramic X-ray images.
• To address the limitations of CBCT in terms of radiation dose and expense.
• To enhance the accuracy and clinical applicability of 3D dental modeling.

Main Methods:

• The DTR-Net framework operates in both image and geometric spaces.
• An image space approach utilizes a 2D-to-3D generative model guided by a tooth segmentation network.
• A geometric space approach employs an implicit function network to capture complex tooth shapes.

Main Results:

• DTR-Net achieves state-of-the-art performance in quantitative and qualitative 3D tooth model reconstruction.
• The method successfully recovers CBCT image intensities and captures intricate geometric tooth properties.
• Experimental results validate the effectiveness of the dual-space framework.

Conclusions:

• DTR-Net presents a significant advancement in 3D tooth model reconstruction from 2D panoramic X-rays.
• The proposed framework offers a promising, cost-effective, and lower-radiation solution for dental practice.
• This technology has strong potential for widespread application in digital dentistry.