3D Tooth Pose Estimation Using Neighbor-Guided Neural Normalization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for tooth pose estimation in virtual 3D representations are labor-intensive, time-consuming, and prone to errors, especially when dealing with heavily ground teeth, and existing algorithms like PCA are inefficient for performance-sensitive applications requiring rapid results.
Innovation Solution
A computer-implemented method using a trained neural network to estimate tooth poses by segmenting a virtual 3D representation, determining initial poses based on geometric parameters of neighboring teeth, and normalizing the representation for faster and more accurate output.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual annotation of coordinate system axes is used for each tooth representation, then measurement precision is improved, but productivity deteriorates due to labor and time intensity
Solution Approach 1:
The patent replaces manual mechanical annotation processes with an automated neural network system. The neural network takes segmented tooth representations as input and automatically predicts accurate tooth poses, eliminating the need for manual coordinate system annotation while maintaining high measurement precision and significantly improving productivity.
Solution Approach 2:
The patent uses a neural network model trained on manually annotated data to create a computational copy that can automatically generate tooth pose estimations. This copied knowledge allows the system to reproduce accurate poses without requiring manual annotation for each new tooth representation, thus scaling productivity while maintaining precision.
2Productivity
If Principal Component Analysis algorithm is used for tooth pose estimation, then productivity is improved with faster computation, but measurement precision deteriorates due to errors in heavily ground teeth
Solution Approach 1:
The patent replaces the traditional PCA algorithm with a neural network-based system. The neural network processes segmented tooth representations and outputs accurate tooth poses, overcoming the limitations of PCA in handling heavily ground teeth while maintaining fast computation speeds. The neural network's learned features allow it to handle complex tooth geometries that PCA cannot process accurately.
3Ease of operation
If current algorithms are used for tooth pose estimation, then ease of operation is improved with automated processing, but productivity deteriorates due to excessive computational time
Solution Approach 1:
The patent changes the computational parameters and architecture by using a neural network model specifically trained for tooth pose estimation. This involves transforming the input data through segmentation and normalization processes, then feeding it into the neural network which outputs poses in under one second. The parameter changes in the computational approach enable both ease of automated operation and high productivity.
Data Source
AI summary
Embodiments relate to obtaining a virtual 3D representation of a patient's dentition, segmenting the virtual 3D representation to obtain first and second segmented tooth representations of neighboring teeth in the patient's dentition. The method can involve determining an initial tooth pose for the first segmented tooth representation using a geometric parameter of the first segmented tooth representation and a geometric parameter of the second segmented tooth representation. A normalized tooth representation of the first segmented tooth representation is obtained by transforming the first segmented tooth representation using the initial tooth pose for the first segmented tooth representation. The normalized tooth representation is then input into a trained neural network to output a correct tooth pose for the first segmented tooth representation.