3D Denture Modeling From CBCT Without Manual Occlusion Registration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for generating 3D models of dentures in occlusion conditions are time-consuming and prone to errors due to difficulties in separating denture representations in 3D images, especially when dealing with varying materials and unknown occlusal plane orientations, leading to increased costs and reduced patient comfort.
Innovation Solution
A method and system for generating 3D models of dentures in occlusion conditions using cone-beam computed tomography (CBCT) that classifies voxels based on neighboring characteristics, employing algorithms such as watershed and deep neural networks to automatically separate denture surfaces without requiring knowledge of the occlusal plane orientation, ensuring accurate and efficient mesh registration.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If conventional thresholding methods are used to separate denture representations in CBCT images, then the process is simple, but the mesh extraction accuracy deteriorates due to difficulty in separating dentures and unknown occlusal plane orientations
Solution Approach 1:
The patent divides the denture separation problem into multiple processing stages: initial thresholding to obtain binary images, followed by contour detection to identify tooth regions, and finally iterative classification algorithms to separate individual dentures. This multi-stage segmentation approach maintains simplicity while improving accuracy by breaking down the complex separation task into manageable steps.
Solution Approach 2:
The patent performs preliminary actions by first obtaining binary images through thresholding before attempting separation, and by detecting contours to identify tooth regions prior to classification. These preliminary steps prepare the data in advance, making the subsequent separation process more accurate without requiring complex algorithms from the start.
2Manufacturing precision
If multiple separate scans are performed for each denture, then mesh extraction accuracy improves, but chair time and production costs increase
Solution Approach 1:
The patent merges multiple denture scans into a single combined CBCT image, allowing all dentures to be captured simultaneously in one scanning operation. This combining approach maintains mesh extraction accuracy through subsequent computational separation while dramatically improving productivity by eliminating the need for multiple separate scanning sessions.
Solution Approach 2:
The patent creates digital copies of the combined scan data and processes them through iterative classification algorithms to generate separate 3D models for each denture. This copying and computational separation approach allows accurate individual mesh extraction from a single combined scan, avoiding the need for multiple physical scans.
3Productivity
If desktop scanners are used to scan dentures, then chair time is reduced, but the process becomes complicated requiring multiple acquisitions and registrations
Solution Approach 1:
The patent extracts the essential separation function from the scanning process itself, performing denture separation through computational algorithms rather than through complex mechanical scanning procedures. This extraction of the separation task to the digital domain simplifies the overall process by eliminating the need for multiple physical acquisitions and manual registrations.
Solution Approach 2:
The patent replaces complex mechanical scanning and manual registration procedures with automated computational algorithms. The iterative classification and contour detection algorithms automatically separate and register dentures in the digital domain, substituting mechanical complexity with computational efficiency.
4Manufacturing precision
If manual mesh registration is performed, then alignment accuracy improves, but time consumption and chair time increase
Solution Approach 1:
The patent implements self-service by enabling the system to automatically perform mesh registration through iterative classification algorithms that use neighboring voxel characteristics. The system registers dentures autonomously without requiring manual intervention, maintaining alignment accuracy through computational methods while eliminating time consumption associated with manual registration.
Solution Approach 2:
The patent employs feedback mechanisms through iterative classification where the algorithm continuously refines denture separation based on neighboring voxel characteristics and previously classified regions. This iterative feedback process automatically achieves accurate alignment and registration without manual intervention, maintaining precision while reducing time consumption.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Enables fast and reliable generation of precise 3D models of dentures, reducing chair time and production costs while improving patient comfort by eliminating the need for manual mesh registration and enhancing mesh extraction accuracy.
Implementation Method 1
x-ray scanning the first and the second dentures in occlusion conditions to obtain a set of data
Implementation Method 2
combining these images to obtain a 3-D reconstruction showing the dentition of the jaw and other facial features for a patient
Data Source
Figure 1a~1b
Figure 1c~2
Figure 3
AI summary
According to some embodiments of the invention, it is provided a method of generating a 3D model of a first denture and of a second denture comprising: x-ray scanning the first and the second dentures in occlusion conditions to obtain a set of data; generating, from the obtained set of data, a 3D image of the first and the second dentures in occlusion conditions, the generated 3D image comprising voxels; classifying each of voxels of the generated 3D image as belonging to the first denture, the second dentures, or space between the first and the second dentures; and generating, from the classified voxels, a 3D model of the first denture and of the second denture, wherein classifying a voxel of the generated 3D image is based on characteristics of neighboring voxels in the generated 3D image, according to an iterative analysis of voxels from neighbor to neighbor.