Marching Surfaces: Isosurface Approximation using G$^1$ Multi-Sided Surfaces

TL;DR

This paper introduces a novel G^1 multi-sided surface approach for isosurface approximation from 3D scalar fields, offering improved continuity, data efficiency, and visual fidelity over traditional methods like Marching Cubes.

Contribution

It presents a new G^1 multi-sided patch interpolation scheme that enhances isosurface approximation, continuity, and data compression for efficient rendering.

Findings

Provides better surface continuity than traditional methods

Achieves data compression suitable for mobile and web applications

Outperforms Marching Cubes in visual fidelity and efficiency

Abstract

Marching surfaces is a method for isosurface extraction and approximation based on a $G^1$ multi-sided patch interpolation scheme. Given a 3D grid of scalar values, an underlying curve network is formed using second order information and cubic Hermite splines. Circular arc fitting defines the tangent vectors for the Hermite curves at specified isovalues. Once the boundary curve network is formed, a loop of curves is determined for each grid cell and then interpolated with multi-sided surface patches, which are $G^1$ continuous at the joins. The data economy of the method and its continuity preserving properties provide an effective compression scheme, ideal for indirect volume rendering on mobile devices, or collaborating on the Internet, while enhancing visual fidelity. The use of multi-sided patches enables a more natural way to approximate the isosurfaces than using a fixed number of sides or polygons as is proposed in the literature. This assertion is supported with comparisons to the traditional Marching Cubes algorithm and other $G^1$ methods.