Automated Tooth Type Recognition from 3D Scans

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing digital workflows for intra-oral scanners require human input for tooth labeling, which can lead to errors due to mislabeling.

Innovation Solution

An automated method for identifying and labeling teeth using segmented digital 3D models, where distinct features are computed and aggregated to classify tooth types, and tooth widths are used to identify types and detect interstices.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If automated tooth identification is implemented, then labeling accuracy is improved, but system complexity increases

Engineering Contradiction:
Improvelabeling accuracyVSAvoidsystem complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The tooth identification system segments the 3D dental model into individual teeth based on detected interstices, then processes each tooth independently through feature extraction and classification. This segmentation allows complex identification to be broken down into manageable steps: first detecting interstices between teeth, then extracting features from each segmented tooth, and finally classifying each tooth type separately, thereby reducing overall system complexity while maintaining high accuracy.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The system introduces an intermediary feature aggregation step that computes a summary feature representation from the detailed 3D tooth geometry. This intermediary representation (such as shape descriptors or statistical moments) serves as a bridge between the complex raw 3D data and the final tooth type classification, simplifying the comparison process while preserving essential identifying characteristics.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Device complexity

If manual tooth labeling is used, then system complexity is reduced, but labeling errors increase

Engineering Contradiction:
Improvesystem complexityVSAvoidlabeling accuracy
Core Design Contradiction:
Device complexityVSReliability

Solution Approach 1:

The system enables the 3D dental model to identify its own tooth types through automated feature extraction and classification algorithms. The process automatically detects interstices, segments teeth, extracts geometric features, and classifies tooth types without requiring human intervention, allowing the data itself to serve its own identification needs while eliminating human error.

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The patent replaces the mechanical human labeling process with an automated computational system that uses algorithms to detect interstices, extract geometric features, and classify tooth types. This substitution eliminates human subjectivity and error while maintaining the ability to handle complex dental geometries through digital processing.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

3Productivity

If feature aggregation is used for tooth classification, then identification speed is improved, but feature loss may occur

Engineering Contradiction:
Improveidentification speedVSAvoidfeature loss
Core Design Contradiction:
ProductivityVSLoss of information

Solution Approach 1:

The system extracts and analyzes local geometric features at specific locations on the tooth surface (such as curvature, surface area, and shape characteristics) rather than treating the entire tooth as a uniform entity. By preserving and analyzing these local quality variations, the system maintains detailed feature information while still being able to aggregate it into summary representations for rapid classification, thus avoiding feature loss.

Inventive Principle:
Principle #3Local quality

Data Source

PatentEP4552609A1Method for automatic tooth type recognition from 3D scans
Publication Date: 2025.05.14 SOLVENTUM INTELLECTUAL PROPERTIES CO
  • EP4552609A1 patent drawingFigure 1~3
  • EP4552609A1 patent drawingFigure 4~5
  • EP4552609A1 patent drawingFigure 6~7

AI summary

The invention relates to a method (22) for identifying a tooth type by classification, comprising steps executed by a processor (20) of receiving (24) a segmented digital 3D model of teeth, wherein each tooth in the model is digitally separated from other teeth; selecting (26) a digital 3D model of a tooth from the segmented digital 3D model, the selected digital 3D model of a tooth being associated with a plurality of distinct 3D mesh features; computing (90) an aggregation of the plurality of distinct 3D mesh features to generate a single 3D mesh feature describing the selected digital 3D model of the tooth; and identifying (26) a type of the selected digital 3D model of the tooth based upon the aggregation.