AI-based shape generation and registration for dental restorations

This project aims to advance the development of a 3D shape generation software that uses generative adversarial networks to design dental crowns. A current prototype version of the software has been developed that allows generating dental restorations represented as point clouds. While currently generated shapes are coherent with the overall generation context of the crown, which includes nearby and opposing teeth, the current point cloud representation suffers from excessive noise and struggles to cope with partial dental arches comprising only a few teeth. This proposal aims to enhance the prototype generative capacity by replacing point clouds by meshes, by incorporating dental preparations as input to the generation pipeline and by improving segmentation of partial arches through improved 3D shape registration. A complementary goal of this project is to improve training and generation speed through the integration of transformers to replace the current convolutional approach used by the encoder.

Faculty Supervisor:

François Guibault;Farida Cheriet;François Guibault;Vladimir Brailovski

Student:

Partner:

Object Research Systems;iMD Research

Discipline:

Computer science

Sector:

Information and cultural industries; Professional, scientific and technical services

University:

Polytechnique Montréal

Program: