Selective Image Compositing With Minimal 3D Geometry
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Generating a complete 3D model of a physical structure for augmenting or modifying portions of it is resource-intensive and can result in lower visual fidelity due to incomplete contextual information and differences between rendering environments, leading to unrealistic modifications.
Innovation Solution
A method that uses a minimum amount of 3D geometric data, such as block or planar geometry with 3D keypoints, to represent a physical structure, combined with synthetic image data, and applies scene effects to create a photorealistic image by aligning it with the original 2D image's camera intrinsics and lighting conditions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If a complete 3D model is generated for augmenting or modifying portions of a physical structure, then the model provides comprehensive geometric representation, but it consumes excessive processing resources and increases image rendering time
Solution Approach 1:
The patent segments the physical structure into multiple facets, where only the selected facet is modeled with complete 3D geometry while other facets use simplified representations. This allows the system to maintain geometric accuracy for the modified portion while reducing overall processing requirements and rendering time.
2Quantity of substance
If a complete 3D model is generated from external image sources, then spatial data is enhanced, but visual fidelity decreases due to rendering environment differences and incomplete contextual information
Solution Approach 1:
The patent applies different quality levels to different parts of the model. The selected facet that requires modification receives high-quality complete 3D modeling with full contextual information, while other facets use lower-quality simplified representations. This local differentiation maintains visual fidelity where needed while reducing overall processing burden.
3Adaptability or versatility
If modifications are applied to a complete 3D model, then comprehensive structural features are available, but the modifications appear unrealistic due to rendering environment differences
Solution Approach 1:
The patent performs preliminary actions by capturing the original image with contextual information (lighting, camera intrinsics, scene effects) before generating the 3D model. This contextual data is stored and later applied to the rendered modifications, ensuring they appear realistic in the final composite image even though the 3D model itself is simplified.
Data Source
AI summary
Disclosed are techniques for generating a photorealistic image by augmenting or compositing at least a portion of a physical structure (e.g., a house) depicted in a two-dimensional (2D) image with synthetic image data. Additionally, disclosed are techniques for augmenting the depicted physical structure and applying a scene effect to the synthetic image data to create a photorealistic effect.