An intelligent design method for complete denture based on multi-task collaborative deep learning

By using a multi-task collaborative deep learning method, the tooth position and base of the denture are adaptively adjusted, which solves the problems of strong reliance on experience and cumbersome operation in the design of complete dentures, and achieves efficient and stable denture design and wearing effect.

CN120983166BActive Publication Date: 2026-03-03PEKING UNIV SCHOOL OF STOMATOLOGY
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202511103611.3
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2025-08-07
Publication Date
2026-03-03
Estimated Expiration
2045-08-07

AI Technical Summary

Technical Problem

The design of complete dentures suffers from problems such as strong reliance on experience, cumbersome and inefficient operation, and difficulty in determining the tissue buffer area and denture boundary in the design of the denture base, which leads to discomfort and poor stability when wearing dentures.

Method used

Employing a multi-task collaborative deep learning approach, this method learns from expert experience through deep learning networks, adaptively adjusts tooth pose, and generates denture bases to achieve intelligent design of complete dentures, including buffering of edentulous jaw models, filling of undercuts, and segmentation of denture base boundaries.

Benefits of technology

It significantly improves the quality, stability, and efficiency of denture design, enhances wearing comfort and stability, reduces reliance on technicians, and promotes its widespread use in primary healthcare institutions.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN120983166B_ABST
    Figure CN120983166B_ABST
Patent Text Reader

Abstract

The application discloses a kind of full mouth denture intelligent design methods based on multi-task cooperative deep learning, applied to oral prosthetic technical field.It includes the following steps: through edentulous jaw buffer area mask self-encoding network to generate tissue face after buffer edentulous jaw model;Adopting in situ channel direction undercut filling algorithm generates after undercut filling model;Based on multi-rigid body pose prediction network generates full mouth denture tooth arrangement position;Using deep learning model to separate out the base plate boundary from edentulous jaw model;With the constraint condition of base plate boundary and full mouth denture tooth arrangement position, combined with the curvature of generated surface to construct composite loss function, using conditional adversarial generative network generates base plate model.The present application automatically core tooth arrangement step, reduces the error caused by human factor, significantly improves the consistency and reliability of denture design quality, enhances the wearing comfort and long-term stability of final denture, helps the popularization and application of high-quality full mouth denture prosthetic technology in primary medical institutions.
Need to check novelty before this filing date? Find Prior Art

Citation Information

Patent Citations

  • Tooth arrangement method for manufacturing complete dentures

    CN105380723A

  • Digital designing and making method of individual tray without edentulous jaw function pressure

    CN107198586A