Textured 3D Dental Models via Quality-Filtered Texture Atlas
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for generating 3D models of dental structures using intraoral scanners face challenges in accurately representing color information, leading to over-blending, wash-out, and limited color sampling resolution, which affects the identification of fine details and cosmetic outcomes in dental treatments.
Innovation Solution
A method is provided to generate textured 3D models by determining texture values for points on a polygon mesh, filtering frames to select candidate texture values, computing quality factors, and combining them to create a texture atlas that maps to the 3D model, allowing for high-resolution color representation independent of geometric resolution.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If vertex-based triangle mesh coloring techniques are used to provide color information data, then color information can be obtained for the 3D model, but memory and processing power constraints prevent generation of sufficiently high resolution color representation
Solution Approach 1:
The patent segments the color information processing by separating geometric mesh vertices from color sampling points. Instead of requiring one color value per vertex, the system identifies a subset of sampling points on the mesh surface and determines color values only for these points by filtering and combining candidate texture values from multiple frames. This segmentation allows high-resolution color representation without requiring every mesh vertex to have full color data, thereby reducing memory and processing requirements.
Solution Approach 2:
The patent introduces a temporal dimension by utilizing multiple frames captured during the scanning process. Instead of relying solely on spatial resolution of the mesh, the system filters candidate texture values across multiple temporal frames to determine color values for sampling points. This temporal dimension provides redundant color information that can be combined to achieve high color resolution without increasing spatial mesh density, thus avoiding the memory and processing burden of ultra-fine geometric meshes.
2Measurement precision
If brute-force techniques are used to increase mesh resolution for better color sampling, then color resolution improves, but hardware performance requirements become prohibitively high
Solution Approach 1:
The patent applies partial action by determining color values for only a subset of points on the mesh surface rather than for every vertex. The system identifies sampling points that are sufficient to capture the essential color information and texture variations, then filters candidate texture values for these specific points. This partial approach achieves adequate color sampling resolution without the excessive computational cost of processing every mesh vertex, thereby reducing hardware power requirements while maintaining diagnostic and cosmetic quality.
3Manufacturing precision
If high-resolution triangle meshes are generated to capture fine color details, then color information accuracy improves, but processing time and computational resources increase significantly
Solution Approach 1:
The patent performs preliminary action by pre-identifying a subset of sampling points on the mesh surface before determining their color values. The system selects these points based on their significance for capturing color information and texture variations, then filters candidate texture values only for these pre-selected points. This preliminary identification step avoids the computational burden of processing all mesh vertices, thereby improving processing efficiency while maintaining color information accuracy through strategic sampling of critical points.
4Device complexity
If vertex-based coloring with interpolation is used, then processing requirements are reduced, but color sampling resolution is limited to one color per vertex
Solution Approach 1:
The patent introduces an intermediary layer between the mesh geometry and color values by using candidate texture values from multiple frames as intermediate data. Instead of directly interpolating colors from vertices, the system filters these intermediary candidate texture values for selected sampling points by evaluating quality factors and combining the most reliable candidates. This intermediary approach enables higher color sampling resolution than simple vertex interpolation while avoiding the full computational cost of processing every vertex with multiple frames, thus balancing processing requirements with color precision.
Data Source
AI summary
A method is provided for generating a texture for a three-dimensional (3D) model of an oral structure. The method includes providing the 3D model of the oral structure in the form of a polygon mesh, identifying a set of points located on the polygon mesh, and determining, for each respective point in the set of points, a respective texture value. Each respective texture value is determined by identifying a set of frames, filtering the set of frames to identify a subset of frames, determining a set of candidate texture values for the respective texture value, computing, for each respective candidate texture value in the set of candidate texture values, a quality factor, and computing the respective texture value for the respective point by combining, based on their respective quality factors, candidate texture values selected from the set of candidate texture values.