3D Digital Structure Registration Using Congruent Feature Sets
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Registering different three-dimensional digital structures into a common coordinate system is challenging due to variations in scan data from different modalities and the presence of noise, especially for physiological objects lacking distinct geometrical features.
Innovation Solution
A method that determines structural features using predefined criteria, identifies congruent feature sets, and applies a rigid spatial transformation through clustering to align the digital structures, enabling automatic registration without prior knowledge of the structures.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional registration methods are used to align three-dimensional digital structures, then registration accuracy can be achieved, but the computational complexity and time consumption increase significantly
Solution Approach 1:
The method segments the three-dimensional digital structures into multiple n-gons with n≥3, where each n-gon is defined by n structural features. This segmentation transforms the complex global registration problem into multiple simpler local correspondence problems between n-gons, reducing computational complexity while maintaining registration accuracy through cumulative feature matching.
Solution Approach 2:
The patent introduces n-gons as intermediary geometric constructs that mediate between the complex three-dimensional structures and the registration algorithm. These n-gons serve as simplified representatives that capture essential structural characteristics, enabling efficient computation of rigid spatial transformations without requiring direct processing of the entire complex structures.
2Measurement precision
If detailed structural analysis is performed to improve registration precision, then alignment accuracy increases, but the processing time increases
Solution Approach 1:
The method performs partial structural analysis by selecting only n structural features (forming n-gons) that are sufficient to define the geometric characteristics needed for registration. This partial action approach achieves adequate alignment accuracy without the excessive processing time required for analyzing all possible structural details of the three-dimensional structures.
Solution Approach 2:
The patent changes the parameter representation from complete three-dimensional structural data to simplified n-gon definitions with n≥3. By transforming the problem parameters from detailed geometric models to essential feature sets, the method reduces processing time while maintaining sufficient alignment accuracy for medical applications.
3Measurement precision
If manual registration methods are used to ensure accuracy for complex shapes, then registration precision improves, but automation level decreases
Solution Approach 1:
The method enables the registration system to perform self-service by automatically identifying corresponding n-gons between different three-dimensional structures and computing the optimal rigid spatial transformation. The algorithm autonomously processes the feature matching and transformation calculation without requiring manual intervention, achieving both high automation level and registration precision through systematic computational procedures.
Solution Approach 2:
The patent implements feedback mechanisms where the registration algorithm iteratively refines the alignment by evaluating the correspondence between n-gons from different structures. The system uses the identified congruent n-gons as feedback to adjust and optimize the rigid spatial transformation, ensuring accurate registration while maintaining full automation throughout the process.
4Reliability
If comprehensive feature sets are extracted to handle noise in scan data, then registration reliability improves, but computational effort increases
Solution Approach 1:
The method applies local quality by focusing computational resources on extracting and matching only the essential n structural features that define the n-gons, rather than processing all possible features throughout the entire structure. This localized feature extraction approach maintains registration reliability by capturing critical geometric information while reducing computational effort through selective processing.
Data Source
AI summary
Disclosed is a computer-implemented method for registering a first and a second three-dimensional digital structure. The first and second three-dimensional digital structures are received. First structural features of the first three-dimensional digital structure and second structural features of the second three-dimensional digital structure are determined. A first plurality of first feature sets of first structural features and a second plurality of second feature sets of second structural features are defined. Pairs of first and second feature sets describing first and second n-gons satisfying a congruence criterion are identified. To each pair of first and second feature set a point in a transformation space is assigned. A rigid spatial transformation for registering the first three-dimensional digital structure with the second three-dimensional digital structure is determined using a clustering of the points assigned in the transformation space. The first and second three-dimensional digital structure are registered using the determined rigid spatial transformation.