3D Human Body Model Reconstruction Using Point-to-Surface Distance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional 3D human body modeling using depth data points from a single viewpoint is inaccurate due to the reliance on point-to-point distance calculations, which become unreliable when vertex resolution is reduced, leading to incomplete and distorted reconstructions.
Innovation Solution
The method calculates point-to-surface distances between depth data points and the surface of a deformed 3D human body model, using triangular faces, edges, and vertices to ensure accurate reconstruction even at reduced vertex resolution, by determining the minimum distance from each data point to the surface, thereby refining the model.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If point-to-point distance calculation is used to reconstruct 3D human body model, then computational cost is reduced, but reconstruction accuracy deteriorates when vertex resolution is reduced
Solution Approach 1:
The patent introduces triangular faces as an intermediary element between depth data points and vertices. Instead of directly calculating point-to-vertex distances, the method calculates point-to-triangle distances using barycentric coordinates, which serves as a more accurate intermediary measurement that maintains reconstruction accuracy even when vertex resolution is reduced.
Solution Approach 2:
The patent transitions from point-to-point distance measurement (1D/0D) to point-to-surface distance measurement (3D). By calculating the perpendicular distance from depth data points to triangular faces on the mesh surface, the method incorporates spatial dimensionality information that point-to-vertex distances lack, thereby improving accuracy without proportionally increasing computational cost.
2Speed
If vertex resolution on deformed 3D human model is reduced, then rendering performance is improved, but reconstruction accuracy deteriorates
Solution Approach 1:
The patent replaces the mechanical/geometric point-to-vertex distance calculation system with a point-to-triangle surface distance calculation system. This substitution uses barycentric coordinate interpolation to determine distances to triangular faces, providing more accurate measurements that are less sensitive to vertex resolution reductions, thereby maintaining reconstruction accuracy while allowing lower vertex counts for better rendering performance.
3Device complexity
If point-to-point distance metric is used, then calculation simplicity is maintained, but measurement precision deteriorates for different vertex resolutions
Solution Approach 1:
The patent changes the measurement parameter from point-to-vertex distance to point-to-triangle distance using barycentric coordinates. This parameter change transforms the distance calculation into a more robust metric that accounts for the local geometry of the mesh surface, providing consistent measurement precision across different vertex resolutions while maintaining reasonable computational complexity through efficient barycentric coordinate calculations.
Data Source
AI summary
Virtual reality-based apparatus that includes a memory, a depth sensor and circuitry. The depth sensor captures a plurality of depth data points of a human subject from a single viewpoint. The memory stores a deformed three-dimensional (3D) human body model. The circuitry calculates first distances from the depth data points to a plurality of triangular faces. The circuitry calculates second distances from the depth data points to a plurality of edges. The circuitry further calculates third distances from the depth data points to a plurality of vertices. The circuitry further determines minimum distances, among the calculated first distances, the calculated second distances, and the calculated third distances, as point-to-surface distance to reconstruct a 3D human body model with high accuracy.