Oral 3D Scan Color Filtering for Foreign Data Removal
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing three-dimensional scanning technologies face inefficiencies in dental applications due to the manual removal of foreign substances and user hands from oral cavity scans, requiring significant post-correction time and resources.
Innovation Solution
An image filtering method that utilizes color information to automatically identify and exclude non-oral cavity data by comparing pixel colors with a reference range, either user-defined or learned, during the conversion from two-dimensional to three-dimensional volume data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple images are synthesized to improve image quality, then image quality is improved, but processing time increases
Solution Approach 1:
The patent applies preliminary action by pre-calculating and storing filter kernels for different motion vectors and search ranges before actual image processing. When synthesizing multiple images, these pre-prepared kernels enable faster filtering operations, reducing the time penalty associated with processing multiple images while maintaining improved image quality through proper motion compensation.
Solution Approach 2:
The patent implements dynamics by adaptively adjusting the search range and motion vector precision based on the specific characteristics of each image block and the desired output quality. This dynamic approach allows the system to use more computational resources for critical regions requiring high quality synthesis while using fewer resources for less critical areas, thereby optimizing the trade-off between image quality and processing time.
2Measurement precision
If a large search range is used in motion estimation, then motion estimation accuracy is improved, but computational complexity increases
Solution Approach 1:
The patent applies segmentation by dividing the search space into multiple regions or stages, where different search strategies are applied to different areas. Instead of uniformly searching the entire large search range, the method segments the problem to focus computational effort on regions with higher probability of containing the optimal motion vector, thereby improving accuracy while reducing overall computational complexity.
Solution Approach 2:
The patent implements local quality by applying different levels of search thoroughness to different spatial regions of the image. Critical regions with complex motion patterns receive more intensive local search with larger effective search ranges, while simpler regions use coarser search strategies. This localized approach maintains high motion estimation accuracy where needed while reducing computational complexity in less demanding areas.
Data Source
Figure 1
Figure 2
Figure 3~4
AI summary
An image filtering method according to the present disclosure comprises: acquiring a two-dimensional image through a scanner, and acquiring color information of at least part of data of the two-dimensional image, and then determining whether the acquired color information is included in a reference color range. When the acquired color information is included in the reference color range, image data remaining after deleting corresponding data from the two-dimensional image is converted into three-dimensional volume data. Meanwhile, when the reference color range is determined, a reference color may be pre-configured data or may be predetermined by a user, or learning for defining a reference color range may be performed through image data acquired by repetitively inputting reference images. By using the image filtering method, a three-dimensional scanner user can minimize post-correction work after scanning and acquire a precise data result value for an interior of the oral cavity, whereby data reliability is enhanced.