3D Avatar Body Boundary Detection for Garment Fitting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing 3D garment simulation systems struggle to accurately modify garments for target avatars with different poses, sizes, and mesh topologies, leading to inefficiencies in dressing the target avatar with the original garment.
Innovation Solution
A method to determine boundary surfaces of body parts on a 3D avatar by generating curvature and thickness information, using Gaussian curvature and predetermined criteria to identify key points, and adjusting garment sizes based on these boundaries.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If curvature information is generated for all points on the 3D avatar surface, then boundary detection precision is improved, but computational complexity increases
Solution Approach 1:
The patent extracts only the necessary curvature information at specific critical points (saddle points, maxima, minima) rather than processing all points on the surface. This selective extraction maintains boundary detection precision while significantly reducing computational complexity by focusing calculations only where geometric features matter for garment boundary identification.
Solution Approach 2:
The method applies different processing quality to different regions of the 3D avatar surface. Instead of uniform processing, it identifies and processes only the local regions with specific curvature characteristics (saddle points, maxima, minima) that are relevant for boundary detection, while skipping uniform or non-critical regions to reduce overall computational load.
2Manufacturing precision
If multiple boundary surfaces are determined based on curvature and thickness criteria, then garment fitting accuracy is improved, but processing time increases
Solution Approach 1:
The patent performs preliminary identification of critical points (saddle points, maxima, minima) and pre-determination of boundary surfaces before the actual garment simulation and fitting process. This preliminary action establishes the boundary framework in advance, reducing processing time during subsequent garment draping and fitting calculations while maintaining high fitting accuracy.
Solution Approach 2:
The boundary detection process is segmented into distinct stages: first identifying critical points based on curvature, then filtering by thickness criteria, and finally determining boundary surfaces. This segmentation allows each stage to be optimized independently and enables parallel processing of different body regions, reducing overall processing time while maintaining comprehensive boundary detection accuracy.
3Adaptability or versatility
If the 3D avatar mesh is divided into multiple segments based on boundary surfaces, then garment customization capability is improved, but mesh topology complexity increases
Solution Approach 1:
The patent segments the 3D avatar mesh into anatomically meaningful regions (torso, limbs, head) based on curvature-based boundary surfaces. This segmentation enables customized garment design for each body part while maintaining a relatively simple base mesh topology. The segmentation is achieved through mathematical boundary detection rather than complex mesh editing, preserving adaptability without proportionally increasing topology complexity.
Solution Approach 2:
The method applies different mesh processing quality to different body regions based on their functional requirements for garment customization. Critical regions for fitting (such as torso and limbs) receive detailed boundary definition, while less critical regions use simplified processing. This localized approach enables comprehensive customization capability while minimizing the overall complexity increase across the entire mesh topology.
Data Source
AI summary
According to an example embodiment, a method of determining a body part boundary surface of a three-dimensional (3D) avatar includes: generating curvature information including a curvature of each of a plurality of points present on a surface of the 3D avatar; determining, from among the plurality of points, first points of which a curvature satisfies a first predetermined criterion based on the curvature information; determining second points from among the first points based on whether a thickness corresponding to each of the first points determined based on thickness information of the 3D avatar satisfies a second predetermined criterion; and determining a plurality of body part boundary surfaces based on the second points.