3D Object Mesh Texturing Using Canonical Views
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Users often face disappointment when receiving items purchased online, as they may not match the envisioned size or style due to lack of photorealistic representations, necessitating improved three-dimensional modeling for better online shopping experiences.
Innovation Solution
A two-stage approach for generating photorealistic three-dimensional representations of objects using canonical views, where a subset of views is used to texture most of the object mesh structure, followed by additional views to ensure seamless transitions and complete texturing, minimizing energy functions to optimize photorealism.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a three-dimensional model is provided for online shopping, then photorealistic representation is improved, but file size increases
Solution Approach 1:
The patent segments the 3D model into a mesh structure with discrete elements, allowing selective texturing of only those mesh elements that correspond to visible surfaces in captured images. This segmentation enables the system to apply detailed photorealistic textures only where needed rather than across the entire model, reducing overall file size while maintaining photorealistic representation for visible areas.
Solution Approach 2:
The patent applies local quality by determining which specific mesh elements require texturing based on their visibility in captured images. Only mesh elements that are visible in at least one captured image are textured with high-quality photorealistic data, while hidden elements remain untextured. This localized approach optimizes the balance between photorealistic representation and file size by concentrating detail where it matters most.
2Measurement precision
If multiple views are captured to improve photorealism, then texturing quality improves, but processing complexity increases
Solution Approach 1:
The patent applies preliminary action by capturing multiple views of the object before creating the 3D model. These pre-captured images are stored and later used to texture the mesh elements. By having the images ready in advance, the system can efficiently map them to corresponding mesh elements without requiring complex real-time processing, thus improving texturing quality while managing processing complexity.
Solution Approach 2:
The patent uses copying by projecting 2D images onto 3D mesh elements. The captured images are copied and mapped to the corresponding mesh elements in the 3D model, transferring photorealistic surface information from the real object to the digital representation. This copying approach simplifies the texturing process compared to generating textures procedurally, as it directly transfers actual visual data.
3Productivity
If canonical views are used to texture most mesh elements, then texturing efficiency improves, but transition seams may appear
Solution Approach 1:
The patent applies the taking out principle by extracting and treating the transition regions separately from the main mesh elements. After using canonical views to texture most mesh elements efficiently, the system identifies and extracts the band regions where transitions occur, then applies additional views specifically to these extracted regions to ensure seamless blending. This separation allows efficient bulk texturing while maintaining high-quality transitions.
Solution Approach 2:
The patent applies local quality by applying different texturing strategies to different regions of the mesh. Canonical views are used for the majority of mesh elements where efficiency is paramount, while additional views are applied locally to transition regions where smoothness is critical. This region-specific approach optimizes both texturing efficiency and transition smoothness by matching the processing method to the local requirements of each mesh region.
Data Source
AI summary
Images of various views of objects can be captured. An object mesh structure can be created based at least in part on the object images. The object mesh structure represents the three-dimensional shape of the object and includes a mesh with mesh elements. The mesh elements are assigned views first from a subset of views to texture large contiguous portions of the object from relatively few views. Portions that are not textured from the subset views are textured using the full set of views, such that all mesh elements are assigned views. The views first assigned from the subset of views and the views then assigned from the full plurality of views can be packaged into a texture atlas. These texture atlas views can be packaged with mapping data to map the texture atlas views to their corresponding mesh elements. The texture atlas and the object mesh structure can be sent to a client device to render a representation of the object. The representation can be manipulated on the client device in an augmented reality setting.