Virtual Model Artifact Removal for Scanned Assets
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The mass production of high-fidelity virtual representations of real-world objects is challenging due to the inclusion of undesirable artifacts from neighboring objects or scanning errors, making manual correction unfeasible for large numbers of objects and limiting the seamless integration of virtual and real-world experiences.
Innovation Solution
The implementation of a system that uses separation and stitching algorithms, combined with color classification and feature vector generation, to automatically isolate and remove non-article vertices from virtual models, facilitating the creation of high-fidelity virtual representations with reduced dependency on manual correction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If manual correction is used to remove artifacts from scanned objects, then the quality of virtual representations is improved, but the time and cost increase significantly making it unfeasible for large numbers of objects
Solution Approach 1:
The system enables virtual representations to correct their own artifacts through automated processing. The computer system automatically identifies and removes artifacts from scan data without requiring manual intervention, allowing the scanning process to self-correct and produce high-quality virtual models efficiently
Solution Approach 2:
The patent replaces the manual mechanical correction process with an automated computer-based system. Instead of artists manually examining and correcting scan data, a computer system with automated algorithms processes the scan data to identify and remove artifacts, substituting human labor with computational processing
2Productivity
If automated systems are used to scan objects, then the productivity is improved, but the ability to isolate the real-world object from surrounding objects and anticipate manual corrections is insufficient
Solution Approach 1:
The system applies different processing treatments to different portions of the scan data based on their characteristics. It identifies specific regions containing artifacts versus regions containing the target object, and applies selective processing to remove artifacts while preserving the integrity of the target object's virtual representation
Solution Approach 2:
The automated system adjusts processing parameters to optimize both isolation accuracy and artifact removal. By dynamically modifying processing parameters based on the specific characteristics of each scan, the system maintains high reliability in object isolation while processing large numbers of objects efficiently
Data Source
AI summary
Various of the disclosed embodiments present systems and methods for distinguishing portions of a virtual model associated with a clothing article from portions of the virtual model not associated with the clothing article. Some embodiments facilitate quick and effective separation by employing a feature vector structure conducive to separation by a linear classifier. Such efficient separation may be especially beneficial in applications requiring the rapid scanning of large quantities of clothing while retaining high-fidelity representations of the clothing's geometry. Some embodiments further accommodate artist participation in the filtering process as well as scanning of articles from a variety of orientations and with a variety of supporting structures.