Surface Scan Inference for Full 3D Intra-Oral Reconstruction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing dental imaging methods, such as physical impressions and intra-oral scans, fail to provide detailed information about the internal structures of teeth and surrounding oral cavity, while CBCT scans, though capable, have lower resolution and health and cost concerns.
Innovation Solution
A machine learning inference model is trained using co-registered three-dimensional image data to map from surface representations of intra-oral tissue to generate a full three-dimensional representation of internal structures, including teeth roots and surrounding tissues, without the need for CBCT scans.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If CBCT scans are used to obtain internal three-dimensional structure information, then measurement precision is improved, but object-affected harmful factors increase due to X-ray exposure and cost
Solution Approach 1:
The patent creates a virtual copy (three-dimensional virtual model) of the internal oral cavity structures by training a machine learning model on CBCT scan data. Once trained, the model generates accurate internal structure representations from standard intra-oral scan data, eliminating the need for actual CBCT X-ray exposure in routine cases. This copying approach preserves measurement precision while removing harmful X-ray exposure.
Solution Approach 2:
The patent performs preliminary action by collecting and processing CBCT scan data during the training phase to build a machine learning model. The model learns the relationship between external surface geometry and internal structures in advance. During actual orthodontic planning, the pre-trained model can generate internal structure predictions without requiring new CBCT scans, thus avoiding X-ray exposure while maintaining accuracy.
2Object-affected harmful factors
If physical dental impressions or intra-oral scans are used, then object-affected harmful factors are reduced, but measurement precision deteriorates due to lack of internal structure information
Solution Approach 1:
The patent introduces a machine learning inference model as an intermediary between standard intra-oral scan data and internal structure information. The model acts as a mediator that translates surface geometry data into predictions of internal structures, enabling the system to obtain internal structure information without direct X-ray exposure while maintaining measurement precision.
Solution Approach 2:
The patent replaces the mechanical/radiographic system (CBCT X-ray scanning) with a computational system (machine learning model). Instead of using physical X-rays to capture internal structures, the system uses a trained neural network that processes surface scan data and generates internal structure predictions, substituting one technical approach with another that eliminates harmful effects.
3Object-affected harmful factors
If heuristics are used instead of CBCT, then object-affected harmful factors are reduced, but measurement precision deteriorates due to reliance on orthodontist expertise
Solution Approach 1:
The patent enables the system to serve itself by creating a machine learning model that automatically predicts internal structures from surface scans. The model, trained on expert CBCT data, encapsulates orthodontic expertise and applies it consistently without requiring individual orthodontist judgment for each case. This self-service approach maintains measurement precision while eliminating the need for harmful CBCT exposure.
Data Source
AI summary
The invention relates to the creation and/or use of an inference model that maps between a three-dimensional image data of an intra-oral tissue structure, e.g. shape and size of teeth roots, and a three-dimensional representation of an exposed surface of dental features of the intra-oral tissue structure, e.g. teeth crowns. In particular, the present invention proposes a method for image processing, using an inference model, a computer-readable surface representation of an individual's intra-oral tissue to provide a full three-dimensional representation of the intra-oral tissue. The present invention also proposes a method for creating the inference model.