Shape-Based Registration for Non-Rigid 3D Models with Large Holes
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional 3D object capture technologies are time-consuming and result in incomplete or noisy 3D models due to occlusions and partial scans, requiring manual processing to create fully closed and smooth models for applications like 3D printing and virtual reality.
Innovation Solution
A computerized method and system for generating closed-form 3D models from partial and noisy scans using shape-based registration, which includes a rough match, deformation graph refinement, and reshaping to fill holes in the scans, leveraging algorithms like Coherent Point Drift and Thin-Plate Spline for non-rigid object registration.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional 3D object capture technology is used to scan objects, then 3D models can be created, but the models are incomplete with holes and gaps due to occlusions and partial scans
Solution Approach 1:
The patent creates a complete 3D model by copying and deforming a reference model to match the partial scan data. The reference model serves as a template that is warped and adjusted to fit the incomplete scan, effectively copying the missing geometry from the reference to fill holes and gaps in the captured data.
Solution Approach 2:
The patent performs preliminary actions by pre-processing the partial scan data through hole filling and surface reconstruction before final model generation. The system prepares the incomplete scan data by estimating missing surfaces and smoothing irregularities, creating a more complete input for subsequent registration and deformation steps.
2Productivity
If traditional 3D scanning is performed on non-rigid objects, then object geometry can be captured, but the process is time-consuming and requires manual processing
Solution Approach 1:
The patent implements self-service by enabling the system to automatically process and correct scan data without human intervention. The automated pipeline includes hole filling, surface reconstruction, model registration, and deformation steps that all execute autonomously, eliminating the need for manual CAD processing and significantly reducing production time.
Solution Approach 2:
The patent utilizes parameter changes in the deformation process to automatically adapt the reference model to the partial scan. By adjusting deformation parameters and optimization criteria, the system automatically finds the best match between reference and scan data, replacing time-consuming manual adjustments with automated parameter optimization.
3Measurement precision
If scans are captured and registered via SLAM to create a 3D model, then spatial mapping can be achieved, but the resulting model is noisy with irregular surfaces
Solution Approach 1:
The patent extracts and separates the noise and irregularities from the scan data through filtering and surface reconstruction operations. By extracting only the essential geometric features and discarding noisy details, the system produces a cleaner, smoother surface that retains the essential object shape while removing artifacts from the scanning process.
Solution Approach 2:
The patent performs preliminary surface reconstruction and noise filtering before final model generation. This pre-processing step smooths irregular surfaces and removes scan artifacts, creating a cleaner input for the registration and deformation processes, thereby improving the overall surface quality of the final model.
4Ease of manufacture
If manual processing is performed using CAD tools to create closed 3D models, then fully formed models can be produced, but the process is complex and difficult to implement
Solution Approach 1:
The patent replaces complex manual CAD operations with automated self-service processing. The system automatically performs hole filling, surface reconstruction, model registration, and deformation correction without requiring user expertise in CAD tools. This automation simplifies the workflow significantly, making 3D model creation accessible to users without specialized training while maintaining high model quality.
Data Source
AI summary
Described herein are methods and systems for closed-form 3D model generation of non-rigid complex objects from scans with large holes. A computing device receives (i) a partial scan of a non-rigid complex object captured by a sensor coupled to the computing device; (ii) a partial 3D model corresponding to the object, and (iii) a whole 3D model corresponding to the object, wherein the partial 3D scan and the partial 3D model each includes one or more large holes. The device performs a rough match on the partial 3D model and changes the whole 3D model using the rough match to generate a deformed 3D model. The device refines the deformed 3D model using a deformation graph, reshapes the refined deformed 3D model to have greater detail, and adjusts the whole 3D model according to the reshaped 3D model to generate a closed-form 3D model that closes holes in the scan.