3D Model Verification Using Rendered Image Difference Comparison
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Verifying differences between a 3D model and the subject by individually comparing captured images with the model is time-consuming.
Innovation Solution
An information processing system that generates a 3D model from multi-viewpoint images and displays information comparing the captured images with rendered images of the model, allowing for direct comparison and difference extraction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If individual comparison of captured images with 3D model is performed, then verification accuracy is improved, but verification time increases
Solution Approach 1:
The patent segments the verification process by generating multiple virtual viewpoint images from the 3D model at different camera positions and angles. These virtual images are then individually compared with corresponding captured images to identify differences. This segmentation allows systematic verification while reducing overall time by parallelizing the comparison process across multiple viewpoints rather than requiring sequential individual comparison.
Solution Approach 2:
The patent creates virtual copies of the 3D model from multiple viewpoints to serve as reference images for comparison. These virtual viewpoint images are generated by rendering the 3D model at various camera positions, effectively creating synthetic copies that can be compared with actual captured images. This copying approach enables accurate verification without requiring time-consuming manual inspection of each individual image pair.
2Ease of manufacture
If 3D model generation is performed from captured images, then model creation is achieved, but difference between model and subject increases
Solution Approach 1:
The patent implements a feedback mechanism where the generated 3D model is compared back against the original captured images through the virtual viewpoint comparison process. Differences identified during this verification stage are fed back to correct and refine the 3D model, improving its accuracy. This feedback loop allows the model to be iteratively improved based on actual image data, resolving the contradiction between ease of creation and precision.
Solution Approach 2:
The patent performs preliminary generation of virtual viewpoint images from the 3D model before final verification. This preliminary rendering allows the system to prepare multiple comparison images in advance, enabling efficient and accurate verification of the model against the original captured images. The preliminary action of generating virtual images facilitates subsequent accurate comparison and model refinement.
Data Source
AI summary
An information processing system includes a processor configured to: generate a three-dimensional model of a subject from an image that is captured by photographing the subject; and display information that results from comparing the image that is captured by photographing the subject with an image that is captured by rendering the three-dimensional model of the subject.