Automated 3D Model Generation Using Image Masks and Virtual Capture
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current 3D modeling techniques face challenges such as high costs and time consumption due to manual rendering by professionals, inaccuracies in machine learning algorithms from insufficient training data, and issues with 3D scanners like incorrect feature tracking on reflective surfaces and poor quality images from dirty scanners.
Innovation Solution
A system and method that generate masks of objects using images, simulate artificial 3D capture environments, and blend artificial surfaces with real-world surfaces to create accurate 3D models, allowing for automated generation and display of 3D models with improved fidelity and efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If manual rendering by trained professionals is used to create bespoke 3D models, then high real-world fidelity is achieved, but the process becomes expensive and time-consuming
Solution Approach 1:
The patent uses photogrammetry to create 3D models by capturing multiple photographs of the real object and automatically generating the model from these images. This copying approach replaces manual professional rendering while maintaining high fidelity to the original object, thereby reducing both time and cost without sacrificing quality
Solution Approach 2:
The patent replaces the mechanical/manual process of professional modelers with an automated computational system that uses image processing algorithms and machine learning to generate 3D models automatically from photographs, eliminating the need for manual intervention while preserving accuracy
2Loss of time
If machine learning algorithms are trained on insufficient training data, then processing time is reduced, but accuracy and reliability of the 3D model deteriorate
Solution Approach 1:
The system performs preliminary actions by capturing a comprehensive set of photographs from multiple angles and positions before the 3D modeling process begins. This preliminary data collection ensures sufficient training data is available, allowing the machine learning algorithms to achieve high accuracy without requiring extensive additional processing or data gathering during the modeling phase
3Extent of automation
If 3D scanners are used to capture objects, then automation is improved, but incorrect tracking of feature points on reflective surfaces occurs
Solution Approach 1:
The patent introduces a non-reflective coating or marker as an intermediary on the reflective surface of the object being scanned. This intermediary provides stable, detectable features that the automated scanning system can track accurately, solving the problem of feature point tracking on reflective surfaces while maintaining automation
Solution Approach 2:
The patent applies markers or coatings with distinct color properties to reflective surfaces to create detectable features. These color-coded markers provide high-contrast, easily trackable features for the scanning system, enabling accurate feature point tracking on surfaces that would otherwise be difficult or impossible to track due to reflectivity
4Productivity
If high throughput 3D scanners operate continuously, then productivity increases, but the stage accumulates dirt and grime that produce poor quality images
Solution Approach 1:
The patent creates a virtual copy or digital representation of the clean scanner stage environment and applies it to correct images captured on the dirty actual stage. This allows continuous operation without cleaning while maintaining image quality through digital restoration that removes artifacts caused by dirt and grime on the scanner stage
Data Source
AI summary
Systems and methods including one or more processors and one or more non-transitory storage devices storing computing instructions configured to run on the one or more processors and perform: generating a mask of an object using one or more images; generating a 3D model of the object using the mask of the object; facilitating displaying a 3D display of the object on an electronic device of a user using the 3D model; receiving, from the electronic device of the user, a zoom selection on the 3D display of the object; in response to receiving the zoom selection, facilitating displaying a zoomed 3D display of the object on the electronic device of the user; receiving, from the electronic device of the user, a zoom rotation selection of the object in the zoomed 3D display; and in response to receiving the zoom rotation selection, facilitating rotating the 3D display of the object in the zoomed 3D display on the electronic device of the user. Other embodiments are disclosed herein.