[Paper Review] Computing quasiconformal folds

Di Qiu, Ka-Chun Lam+1 more|arXiv (Cornell University)|Apr 11, 2018

3D Shape Modeling and Analysis30 references4 citations

TL;DR

This paper introduces a novel computational framework for modeling surface folding using quasiconformal geometry, based on solving a linear PDE—the alternating Beltrami equation—with the alternating Beltrami coefficient as a key control parameter. The method enables precise geometric control over folding, solves inverse problems for flat-foldable surfaces from partial data, and supports applications like fold sculpting, Miura-ori pattern generation, and self-occlusion reasoning via automatic unfolding and image in-painting.

ABSTRACT

We propose a novel way of computing surface folding maps via solving a linear PDE. This framework is a generalization to the existing quasiconformal methods and allows manipulation of the geometry of folding. Moreover, the crucial quantity that characterizes the geometry occurs as the coefficient of the equation, namely the Beltrami coefficient. This allows us to solve an inverse problem of parametrizing the folded surface given only partial data but with known folding topology. Various interesting applications such as fold sculpting on 3D models and self-occlusion reasoning are demonstrated to show the effectiveness of our method.

Motivation & Objective

• To develop an intrinsic, geometry-controlled method for modeling surface folding that generalizes conformal parametrization to allow for folding with controlled distortion.
• To solve the inverse problem of reconstructing a flat-foldable surface from partial geometric data and known folding topology, particularly in cases with self-occlusion.
• To enable practical applications such as fold sculpting, texture generation, and occlusion reasoning on folded 3D surfaces.
• To provide a variational and discrete framework for solving the alternating Beltrami equation efficiently via sparse linear systems.
• To demonstrate the method's robustness in handling complex foldings, including multi-folded and cusp-folded surfaces, with automatic unfolding from partial data.

Proposed method

• The folding map is computed as the solution to the alternating Beltrami equation, a linear PDE where the coefficient is the alternating Beltrami coefficient, which encodes local conformal distortion and orientation changes.
• The method uses a quadratic variational formulation of the alternating Beltrami equation, generalizing least-squares conformal parametrization to allow for folding with prescribed distortion.
• Discretization is achieved via a geometric finite element method that respects the coupled nature of the real and imaginary parts of the solution, enabling stable and accurate computation.
• The inverse problem is solved using a reinforcement iteration algorithm that iteratively refines the generalized Beltrami coefficient to recover the unfolded domain from folded data.
• Unfolding is fully automatic and robust, even for complex foldings with multiple layers or cusp-like features, by leveraging the PDE-based inverse map.
• Image in-painting is performed on the recovered unfolded domain using patch-matching algorithms, followed by mapping back to the folded surface for reconstruction.

Experimental results

Research questions

• RQ1 Can quasiconformal geometry be used to model surface folding in a way that provides intrinsic control over the folding geometry?
• RQ2 How can the inverse problem of reconstructing a flat-foldable surface be solved from partial data and known folding topology?
• RQ3 Can the alternating Beltrami equation be effectively discretized and solved as a sparse linear system to enable practical computation on 3D surfaces?
• RQ4 To what extent can the method recover self-occluded regions in folded surfaces using only boundary and singular set data?
• RQ5 Can the framework be extended to generate new Miura-ori patterns and perform fold-like texture synthesis on arbitrary surfaces?

Key findings

• The alternating Beltrami equation provides a stable and geometrically meaningful framework for computing folding maps, with solutions that preserve key topological and geometric features of the folding.
• The method successfully reconstructs the full unfolded domain of a folded surface from only partial boundary data and singular set information, even in cases with complex folding orders.
• The reinforcement iteration algorithm converges empirically to accurate unfoldings, producing regular meshes without unnatural curvatures, even for 2-folded and cusp-folded surfaces.
• The recovered unfolded domains enable effective image in-painting using standard patch-matching techniques, with results closely matching ground truth in both qualitative and quantitative assessments.
• The framework enables new applications such as fold sculpting, texture generation, and Miura-ori pattern design, demonstrating its versatility and robustness on diverse 3D models.
• The method outperforms naive unfolding by automatically inferring folding order and topology, making it suitable for real-world scenarios with missing or occluded data.