Linear Cage Deformer for Stable Multi-Layer Outfit Fitting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for rendering multi-layered clothing on avatars using radial basis functions (RBF) result in distorted shapes and image quality degradation due to compounded distortions as more layers are added, leading to noticeable bulkiness and potential explosion of the outfit.
Innovation Solution
Employing a linear cage deformer technique that adjusts outer cage vertices based on local coordinate frames and preserves original spacing and alignment between inner and outer cages, eliminating extrapolation errors and reducing artifacts.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If radial basis functions (RBF) are used to reshape garment outer cages for multi-layered clothing, then the clothing can be fitted onto the avatar body, but the distortions compound and amplify as more layers are added, causing noticeable bulkiness and potential explosion of the outfit
Solution Approach 1:
The patent segments the deformation process into two distinct parts: (1) a global deformation component that handles overall shape adjustment, and (2) a local deformation component that preserves fine details and spacing. This segmentation prevents the compounding of distortions by treating different aspects of deformation separately rather than applying uniform RBF transformation to all layers.
Solution Approach 2:
The patent applies different deformation strategies to different parts of the garment cage. The local deformation component specifically targets the preservation of spacing and alignment between inner and outer cages, while the global component handles overall fitting. This local quality approach ensures that critical geometric relationships are maintained even as multiple layers are added.
2Ease of operation
If RBF interpolation and extrapolation are used to model continuous spatial data for garment deformation, then the clothing can conform to avatar body shapes, but extrapolation leads to artifacts that diminish image quality
Solution Approach 1:
The patent extracts and removes the problematic extrapolation component from the deformation process. By using a hybrid approach that combines global deformation with local deformation based on original cage geometry, the method eliminates the source of artifacts while preserving the ability to conform garments to avatar bodies.
Solution Approach 2:
The patent employs asymmetric treatment of deformation: the global component handles overall shape adaptation, while the local component uses the original cage geometry to preserve local features. This asymmetric approach allows different parts of the deformation process to serve different purposes, improving overall quality without sacrificing conformability.
3Ease of manufacture
If single layer distortions are applied sequentially to each garment layer using RBF technique, then each layer can be fitted individually, but the distortions compound and get amplified as more layers are added
Solution Approach 1:
The patent performs preliminary action by computing the local deformation component based on the original cage geometry before the distortion compounding occurs. This pre-computed local information is then applied to each layer, preventing the accumulation and amplification of distortions while maintaining individual layer fitting capability.
Solution Approach 2:
The patent introduces an intermediary deformation component that acts as a mediator between the global shape adjustment and the local geometry preservation. This intermediary local deformation component, derived from original cage geometry, prevents distortion compounding while allowing sequential layer fitting to proceed.
Data Source
AI summary
Techniques presented herein provide a way to provide a three-dimensional (3D) avatar with multi-layered clothing. The techniques may include generating an estimated outer cage by layering a plurality of garments onto an avatar body of the 3D avatar in a specific order, generating a target outer cage by laying the specific order with compression. The layering with compression comprises adding a first outer garment over an inner garment previously layered onto the avatar body, the first outer garment being layered based on a rigidity of the first outer garment and an outer layer weight of an outfit formed by the plurality of garments. Then, compression is applied to the estimated outer cage based on using the target outer cage as a compression target to generate a final outer cage.