3D Shape Space Registration via Progressive Correspondence Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Establishing dense correspondences among raw 3D scans with significant shape variability and noise is challenging, requiring manual intervention or strong shape priors, which is time-consuming and costly.
Innovation Solution
A computing system uses a progressive correspondence estimation method, combining a linear model and a nonlinear deformation model to iteratively register unregistered 3D digital shapes to a shape template, enhancing the shape space with each registration.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If non-rigid registration is used to align scans with a template body mesh, then shape correspondence can be established, but manual intervention or strong shape priors are needed when shape variability is large or contains holes and noise
Solution Approach 1:
The system performs preliminary actions by first registering a subset of scans to create an initial shape space, then using this preliminary structure to guide the registration of remaining scans. The progressive approach uses already-registered scans as references for registering new scans, eliminating the need for manual intervention in each step.
Solution Approach 2:
The system enables self-service by using the registered scans themselves as the reference framework for subsequent registrations. The shape space built from a subset of scans automatically serves as the registration target for other scans, making the system self-sufficient without requiring external manual annotation or strong shape priors.
2Measurement precision
If manual annotation of landmark correspondence is performed, then registration accuracy can be improved, but the process is expensive and time-consuming
Solution Approach 1:
The system eliminates manual annotation by using automatically registered scans as the reference framework. The progressive registration process uses geometric and metric information from already-registered scans to automatically establish correspondence for new scans, making the entire system self-sufficient without human intervention.
Solution Approach 2:
The patent replaces the mechanical manual annotation process with an automated computational registration system. Instead of manually marking landmarks, the system uses algorithmic registration based on geometric correspondence and metric preservation to automatically establish landmark correspondence across scans.
3Measurement precision
If a large number of scans are registered to create comprehensive shape space, then shape representation accuracy improves, but the computational complexity and time required increase significantly
Solution Approach 1:
The system segments the registration process into progressive steps, where a subset of scans is first registered to create an initial shape space, then this partial result is used to register remaining scans. This segmentation allows the system to build shape space representation incrementally rather than attempting to process all scans simultaneously, reducing computational complexity at each step.
Solution Approach 2:
The system performs preliminary registration of a subset of scans to establish an initial shape space before proceeding with the registration of remaining scans. This preliminary action creates a foundational reference that simplifies subsequent registration tasks, reducing the overall computational complexity required to process the complete dataset.
4Ease of manufacture
If non-rigid registration is used with limited variations and clean shapes, then registration works well, but it fails when shape variability is large or contains holes and noise
Solution Approach 1:
The progressive registration system uses feedback from already-registered scans to guide the registration of new scans with diverse shapes. The registered scans provide reference information that adapts the registration process to handle varying shapes, holes, and noise, allowing the system to learn from previous successful registrations and apply that knowledge to more challenging cases.
Data Source
AI summary
In some examples, a computing system generates a linear model representing an estimate of a shape space using a first set of registered 3D digital shapes registered to a shape template. The computing system determines a nonlinear deformation model for the shape space using a second set of registered 3D digital shapes registered to the shape template. The computing system creates an initial registration to the shape space for an unregistered shape using the linear model. The computing system predicts an updated registration based on the initial registration using the nonlinear deformation model. In response to determining a shape distance between the updated registration and the unregistered shape being below a threshold value, the computing system adds the updated registration to the first set of registered 3D digital shapes to obtain an updated first set of registered 3D digital shapes.