Efficient C2 Continuous Surface Creation Technique Based on Ordinary Differential Equation

Shaojun Bian, Greg Maguire, Willem Kokke, Lihua You, Jianjun ZhangSymmetry 2019 · DOI

TL;DR

This paper introduces a novel method for creating smooth, continuous 3D surfaces (C2 continuity) using ordinary differential equations (ODEs). This approach significantly reduces model data size and simplifies the creation process by avoiding manual stitching of surface patches, offering more intuitive shape control.

What Problem Does This Paper Address?

Creating detailed 3D models, especially for characters, often involves complex surfaces. Traditional methods like polygon modeling can result in massive data sizes, while patch modeling techniques (like NURBS) require tedious manual effort to ensure smooth transitions (C2 continuity) between different surface patches. This manual stitching is error-prone and time-consuming. Furthermore, existing methods often lack a direct link to underlying physical principles, limiting the realism of deformations. The paper aims to address these issues by providing an efficient, automated, and more physically grounded method for generating C2 continuous surfaces.

Methods

The core of the proposed method is an Ordinary Differential Equation (ODE)-based approach. Specifically, it formulates the surface creation problem as solving a vector-valued sixth-order ODE. This ODE is constrained by boundary conditions that include the shape of boundary curves and their first and second partial derivatives.

Here's a breakdown of the key steps:

  • Problem Transformation: Creating a 3D surface is transformed into finding a solution to a sixth-order ODE.
  • Boundary Constraints: The solution must satisfy specific conditions at the boundaries of the surface patch. These conditions involve the position, tangent (first derivative), and curvature (second derivative) of the boundary curves.
  • Two Approaches for Constraints:
    1. Boundary Curves Only: Two boundary curves are defined, and the tangent and curvature constraints are derived from these curves using specific coefficients.
  • Boundary and Control Curves: Two boundary curves and four additional control curves are used to define the tangent and curvature constraints.
  • ODE Solution: The paper derives an analytical solution for the complementary part of the ODE. This solution, when combined with the boundary constraints, defines the surface patch.
  • C2 Continuity: A key advantage is that by satisfying the position, first, and second derivative constraints at the boundaries, the method naturally ensures C2 continuity between adjacent surface patches, eliminating the need for manual stitching.
  • Shape Control: The shape of the surface can be manipulated by adjusting the parameters of the ODE and the boundary constraints (first and second partial derivatives).

Datasets

Not discussed in the paper. The paper focuses on a new modeling technique and demonstrates its capabilities with generated examples rather than training on specific datasets.

Key Insights and Ideas

  • ODE as a Surface Generator: The paper elegantly frames surface generation as solving a differential equation, moving beyond purely geometric definitions.
  • Natural C2 Continuity: The method inherently enforces C2 continuity by design, directly addressing a major pain point in traditional patch modeling. This is a significant simplification for users.
  • Data Size Reduction: By representing surfaces with boundary curves and derivative information rather than millions of polygon vertices, the method achieves substantial data compression. This is crucial for efficient storage and transmission.
  • Physics-Inspired Control: The ODE formulation hints at a connection to physical processes, suggesting potential for more realistic deformations and animations.
  • Intuitive Shape Manipulation: Deforming surfaces by adjusting derivatives and ODE parameters is presented as more intuitive than manipulating individual polygon vertices.

Key Results

  • C2 Continuity Achieved: The proposed method successfully creates surface patches with C2 continuity, as demonstrated visually and through mathematical formulation.
  • Significant Data Compression: The paper shows that the data size of models created using their approach can be reduced to approximately one-third to one-fifth of the original polygon model size (e.g., a male face model reduced from 4081 vertices to 1467 curve variables).
  • Controllable Surface Shapes: Experiments show that by manipulating shape control parameters and boundary derivatives, a wide range of surface shapes can be generated, including those not achievable with standard tools like Maya's loft operation.
  • Application to Complex Objects: The technique is demonstrated to be applicable to creating complex objects by decomposing them into patches, and even for facial modeling by using curve networks derived from facial action coding systems.

Limitations and Future Work

  • Computational Cost of Solving ODEs: While the paper presents an analytical solution for the complementary part of the ODE, solving complex ODEs can still be computationally intensive, especially for interactive applications. The paper mentions that for complicated surface creation, heavy numerical calculations might still be required.
  • Particular Solution for Sculpting: The paper focuses on the complementary solution for surface creation and manipulation. It notes that using the particular solution of the ODE for "sculpting force" and more advanced surface manipulation will be addressed in future work.
  • Analytical Solution for Specific Cases: The analytical solution derived is for specific conditions (e.g., 4ρλ/η² < 1). The paper mentions that other situations can be treated with similar methodology, but it implies that the full analytical solution might be complex for all parameter ranges.
  • Interactive Real-time Performance: While the method aims for efficiency, achieving real-time interactive performance for highly complex models might still be a challenge.

Future work suggested by the authors includes exploring the use of the particular solution of the ODE for advanced surface manipulation and sculpting, and further optimizing the computational efficiency for interactive applications.

Summary updated

Abstract

In order to reduce the data size and simplify the process of creating characters’ 3D models, a new and interactive ordinary differential equation (ODE)-based C2 continuous surface creation algorithm is introduced in this paper. With this approach, the creation of a three-dimensional surface is transformed into generating two boundary curves plus four control curves and solving a vector-valued sixth order ordinary differential equation subjected to boundary constraints consisting of boundary curves, and first and second partial derivatives at the boundary curves. Unlike the existing patch modeling approaches which require tedious and time-consuming manual operations to stitch two separate patches together to achieve continuity between two stitched patches, the proposed technique maintains the C2 continuity between adjacent surface patches naturally, which avoids manual stitching operations. Besides, compared with polygon surface modeling, our ODE C2 surface creation method can significantly reduce and compress the data size, deform the surface easily by simply changing the first and second partial derivatives, and shape control parameters instead of manipulating loads of polygon points.
BibTeX
@inproceedings{bian2019efficient,
  author = {Shaojun Bian and Greg Maguire and Willem Kokke and Lihua You and Jianjun Zhang},
  title = {{Efficient C2 Continuous Surface Creation Technique Based on Ordinary Differential Equation}},
  booktitle = {Symmetry},
  year = {2019},
  doi = {10.3390/sym12010038},
}
Figure