3D Oral Scan Object Identification Using Relative Position Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing dental scanning technologies struggle with accurately acquiring three-dimensional data of target objects within a mouth due to interference from unnecessary objects like fingers or treatment instruments, requiring manual user intervention for data modification which is burdensome.
Innovation Solution
An identification apparatus and method that utilizes an input unit to receive position data and an arithmetic unit with an estimation model trained to identify and remove unnecessary objects based on their fixed relative positional relationships within the mouth, using machine learning techniques to automate the process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If manual user intervention is used to identify and select three-dimensional data for deletion, then the user can remove unnecessary objects from the scan data, but the operation becomes burdensome and time-consuming
Solution Approach 1:
The system performs automatic identification and classification of objects in the three-dimensional scan data without requiring manual user intervention. The arithmetic unit independently analyzes the scan data, identifies unnecessary objects based on their characteristics and positional relationships, and separates them from target objects, enabling the system to serve itself rather than relying on external user input for this task
Solution Approach 2:
The patent replaces the manual mechanical process of user identification and selection with an automated computational system. The arithmetic unit uses algorithms to automatically detect, classify, and separate objects in the scan data, substituting the mechanical human action of visual inspection and manual selection with an automated digital processing system that analyzes spatial relationships and object characteristics
2Extent of automation
If automatic identification using machine learning is implemented, then the identification speed and automation level increase, but the system complexity increases due to the need for trained estimation models
Solution Approach 1:
The system performs preliminary training of the estimation model using training data that includes three-dimensional scan data and corresponding object labels. This preliminary action of training the model beforehand allows the system to automatically identify and classify objects during actual operation without requiring complex real-time decision-making logic, as the model has already learned the patterns and relationships during the training phase
Solution Approach 2:
The patent utilizes changes in parameters such as position data, shape characteristics, and spatial relationships to differentiate between target objects and unnecessary objects. By monitoring and analyzing these parameter variations, the system can automatically classify objects without requiring overly complex structural designs, relying instead on sophisticated parameter analysis within the estimation model
3Measurement precision
If the system analyzes position data of multiple objects to identify target objects, then the identification accuracy improves, but the computational load increases
Solution Approach 1:
The system segments the analysis process into distinct stages: first identifying all objects in the scan data, then classifying each object based on its position data and spatial relationships with other objects, and finally separating target objects from unnecessary objects. This segmentation allows the computational load to be distributed across multiple simpler processing steps rather than requiring one complex simultaneous analysis of all objects
Data Source
Figure 1
Figure 2
Figure 3
AI summary
An identification apparatus (1) includes an input unit to which position data including position information of each of a plurality of objects is input, and an arithmetic unit (11) that identifies at least one object among the plurality of objects based on the position data input from the input unit and an estimation model (122) that is trained to estimate each of the plurality of objects based at least on the position data of each of the plurality of obj ects, wherein a relative positional relationship of the plurality of objects inside the mouth is fixed.