Mesh Mapping via Curve Approximation and Energy Minimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional surface parameterization techniques fail to accurately map label information between two shape models represented by triangular mesh elements, often resulting in inappropriate boundary mappings and the generation of microscopic errors like reversed triangular mesh elements.
Innovation Solution
A method involving the approximation of region boundaries by curves, movement of nodes based on curve approximation, and energy minimization to maintain shape and area changes, ensuring nodes on boundaries align correctly between spherical surfaces, while preventing the formation of inappropriate elements like reversed parts.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional surface parameterization using spring model energy minimization is used to map nodes to spherical surface, then the mapping process is simple and computationally efficient, but microscopic errors occur at boundaries causing reversed triangular mesh elements and failure to achieve proper bijection
Solution Approach 1:
The patent segments the surface into multiple regions with different energy functions. The first region (source surface) uses a first energy function that prevents reversed elements, while the second region (target spherical surface) uses a second energy function. This segmentation allows each region to be optimized with appropriate constraints, achieving both computational efficiency and boundary precision.
Solution Approach 2:
The patent applies different energy functions to different spatial regions. The first energy function is applied locally to the source surface region to prevent reversed triangular elements, while the second energy function is applied to the target spherical surface region. This local differentiation ensures that each region's specific requirements are met without compromising overall mapping quality.
2Ease of operation
If spring model energy minimization is used without regional constraints, then the overall mapping process is straightforward, but label information boundaries do not align properly between source and target surfaces
Solution Approach 1:
The patent divides the mapping problem into two segmented regions: the source surface region with label information and the target spherical surface region. Each region has its own energy function and constraints. This segmentation enables proper alignment of label boundaries while keeping the overall process manageable through regional optimization.
Solution Approach 2:
The patent implements feedback mechanisms where the energy minimization process iteratively adjusts node positions based on boundary alignment requirements. The first energy function provides feedback to prevent reversed elements, and the second energy function provides feedback to align boundaries, ensuring label information is properly mapped.
3Adaptability or versatility
If conventional energy minimization allows nodes to move freely on spherical surface, then the mapping adapts well to spherical geometry, but triangular mesh elements become reversed and bijection fails
Solution Approach 1:
The patent applies local quality constraints through the first energy function on the source surface region. This energy function specifically prevents reversed triangular mesh elements by constraining node movements locally, while still allowing the overall mapping to adapt to spherical geometry through the second energy function on the target surface.
Solution Approach 2:
The patent applies preliminary anti-action by using the first energy function to preemptively prevent reversed elements before they can form. This constraint is applied during the energy minimization process to counteract the tendency of nodes to move into positions that would create reversed triangles, ensuring bijection reliability throughout the mapping process.
Data Source
AI summary
A mapping method includes; mapping first and second models, each represented by polygonal elements of meshes and including a same number of regions, to a first and a second spherical surfaces, respectively; approximating boundaries of the regions by curves and moving nodes based on the curve approximation; associating the nodes on the boundary of first sphere with points on the boundary of the second sphere; moving the nodes other than the nodes on the boundary by minimizing changes of shapes and areas of the polygonal elements under a constraint that the nodes on the boundary of the first sphere are placed at positions corresponding to the associated points on the second sphere; and calculating a point in the second model for each of the nodes in the first model, from the corresponding node after the movement on the first sphere and corresponding polygonal elements on the second sphere.