3D Object Reconstruction Using Quality Depth Maps From Discrete Views
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing 3D reconstruction techniques, such as SLAM, rely on continuous scanning, which is difficult and time-consuming, especially for mobile devices with complex backgrounds, and are impractical for self-scanning of body parts like feet, leading to inaccurate and inefficient reconstructions.
Innovation Solution
The use of discrete image-based depth maps from multiple perspectives, combined with quality checks and interactive user interfaces, ensures high-quality depth maps are obtained, allowing for accurate 3D reconstructions of asymmetric objects like feet, using fewer images and less computational resources.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If continuous scanning techniques like SLAM are used, then 3D reconstruction can be performed, but the process becomes difficult and time-consuming, especially for mobile devices with complex backgrounds
Solution Approach 1:
The patent replaces continuous scanning with discrete periodic captures from multiple fixed viewpoints. Instead of continuously scanning, the system captures a finite set of images at specific positions and orientations, then reconstructs the 3D model from these discrete views using structured light and depth maps, significantly reducing operational difficulty and time requirements
Solution Approach 2:
The patent segments the scanning process into discrete, manageable steps: capturing images at specific viewpoints, generating depth maps for each view, performing quality checks, and finally assembling the 3D reconstruction. This segmentation transforms the complex continuous scanning task into simpler, more controllable discrete operations
2Adaptability or versatility
If continuous scanning is used, then 3D reconstruction can be performed, but it is impractical for self-scanning of body parts like feet
Solution Approach 1:
The patent enables self-service scanning where users can independently scan their own body parts without requiring professional equipment or expertise. The system provides automated quality checks, guides users through the process, and handles the complex 3D reconstruction automatically, making it practical for everyday self-scanning applications
Solution Approach 2:
The patent adapts the scanning parameters and processing algorithms specifically for body parts like feet, which have asymmetric shapes and specific geometric characteristics. The system adjusts depth map generation, quality check criteria, and reconstruction parameters to optimize for these specific objects, making the process practical for self-scanning
3Manufacturing precision
If existing 3D reconstruction techniques are used, then reconstructions can be generated, but they are inaccurate and inefficient
Solution Approach 1:
The patent implements quality checks that provide feedback on the quality of depth maps and intermediate reconstruction results. The system evaluates metrics such as depth consistency, geometric accuracy, and completeness, and can request additional captures or adjust processing parameters to improve accuracy without significantly reducing efficiency
Solution Approach 2:
The patent replaces traditional mechanical continuous scanning with an optical-based discrete image capture system using structured light and depth sensors. This substitution enables faster, more accurate captures from multiple viewpoints, improving both the accuracy and efficiency of 3D reconstruction
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This method enables fast, efficient, and accurate 3D reconstructions of objects, particularly asymmetric ones, using still images, with improved user convenience and flexibility, enabling applications like custom shoe fitting and additive manufacturing.
Implementation Method 1
obtaining the image data from a depth sensor and/or a LiDAR camera
Data Source
AI summary
Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for reconstructing three-dimensional (3D) objects using depth scans. In one aspect, a method of generating a 3D reconstruction of an object includes obtaining, for each view in a set of views of the object, a quality depth map that represents location coordinates of points on the object from the view, wherein the obtaining comprises. For each view, image data representing one or more images of the object from the view are obtained and one or more depth maps that represent location coordinates of points on the object are obtained. A quality depth map that satisfies one or more quality checks is identified for each view. The 3D reconstruction of the object is generated using the quality depth map for each view in the set of views.