Volumetric Image Generation From Limited Viewpoints
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for generating 3D video from 2D images with limited viewpoints often result in artifacts such as discontinuities, mismatches, and blurring due to the lack of visible data for portions of the subject's body or clothing, leading to unacceptable quality in volumetric images.
Innovation Solution
A system and method that generate an initial volumetric image with higher quality from visible portions and reduce the quality of invisible portions by adjusting resolution, simplifying geometry, changing texture, or combining these methods to match the quality of unseen areas, allowing for customization and integration into environments with specific patterns.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If geometry and texture are generated for invisible portions of the subject, then complete volumetric coverage is achieved, but artifacts such as discontinuities, mismatches, and blurring occur
Solution Approach 1:
The patent applies different quality levels to different regions of the volumetric image. Visible portions maintain high quality with detailed geometry and texture, while invisible portions use simplified geometry and procedural textures. This local differentiation resolves the contradiction by accepting lower quality in specific regions to avoid artifacts while maintaining overall completeness.
Solution Approach 2:
The system changes geometric parameters (complexity, detail level) and texture parameters (resolution, detail) based on visibility information. By dynamically adjusting these parameters between visible and invisible regions, the patent achieves complete coverage without introducing artifacts from attempting uniform high-quality generation throughout.
2Manufacturing precision
If high quality geometry and texture are generated for all portions, then complete coverage is achieved, but computational complexity and processing time increase significantly
Solution Approach 1:
Instead of applying full-quality generation to all portions, the patent applies high-quality processing only where necessary (visible regions) and uses simplified methods for invisible regions. This partial action approach maintains acceptable overall quality while dramatically reducing computational complexity and processing requirements.
3Manufacturing precision
If resolution is maintained uniformly throughout the volumetric image, then image quality is consistent, but file size and storage requirements increase
Solution Approach 1:
The patent implements variable resolution where visible portions maintain high resolution for quality consistency, while invisible portions use lower resolution. This local differentiation maintains perceived quality consistency in visible areas while significantly reducing overall data volume and storage requirements.
Data Source
AI summary
A method for generating a volumetric image of a subject from at least one 2 dimensional image, said at least one 2 dimensional image having a limited number of viewpoints, said volumetric image insertable into an environment.