Pith. sign in

REVIEW 3 major objections 4 minor 1 cited by

A Structure-Preserving Numerical Method for Harmonic Maps Between High-genus Surfaces

T0 review · 3 major / 4 minor · reviewed 2026-08-05 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read For closed surfaces of genus at least two, this paper proves that a discrete harmonic map defined with canonical hyperbolic edge weights exists, is unique within each homotopy class, and is an embedding, and it provides a Riemannian gradien

desk verdict A genuinely useful algorithm for hyperbolic harmonic maps when the input is Delaunay, but the paper's general-case workaround (arbitrary positive weight reset) is outside the theory. read the letter →

arxiv 2509.01256 v1 pith:2YHNYACN submitted 2025-09-01 math.NA cs.NAmath.OC

Zhipeng Zhu , Wai Yeung Lam , Lok Ming Lui This is my paper · ORCID
classification math.NAcs.NAmath.OC MSC 58E2005C1053C4365D18
keywords discreteharmonicmapshyperbolicsurfacescanonicaledgeweightsDelaunaytriangulationhigh-genusRiemanniangradientdescentuniformizationgeometryprocessing
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

High-genus surfaces (genus at least two) carry a canonical hyperbolic metric, and this paper builds a discrete theory of harmonic maps—maps minimizing a total-stretching energy within a chosen homotopy class—between them. The central move is a set of canonical edge weights derived from the hyperbolic metric that generalize the Euclidean cotangent weights; when these weights are positive, the discrete Dirichlet energy has a unique minimizer in every homotopy class, and that minimizer is an embedding. The paper also proves that isometries stay harmonic in this discretization, and gives a Riemannian gradient-descent algorithm—implemented on the universal cover—that converges at least linearly near the solution. This matters practically because earlier discrete harmonic-map pipelines for high-genus surfaces reused Euclidean cotangent weights heuristically, losing these structural guarantees; the new method yields globally valid maps and supports applications such as remeshing without cutting the surface into simply connected pieces.

What carries the argument

The central object is the canonical hyperbolic edge weight, a per-edge scalar constructed from the angles and lengths of the source hyperbolic triangulation (formula (2.8)); it generalizes the Euclidean cotangent weights and carries the structure-preserving properties of the method—positivity (equivalent to the Delaunay condition) yields injectivity, and the weights make discrete isometries harmonic. The second load-bearing mechanism is the Riemannian gradient flow on the product space of lifted vertex positions in the Poincaré disk: the exponential map keeps iterates inside the disk, while the Fuchsian-group constraint on the universal cover enforces the chosen homotopy class on the quotien

What would settle it

Take a genus-2 surface and a congruent target with a known isometry between them; if the algorithm's output does not coincide with that isometry, the harmonicity-of-isometries claim fails. Alternatively, take a deliberately non-Delaunay triangulation, run the algorithm twice with two different positive reset values for the negative weights, and check whether the two outputs are the same embedding; if they differ or one is non-injective, the central theorem does not cover the implemented algorithm.

Watch

Extended reading notes

Core claim

On the paper's own terms, the discovery is that the right discretization of the Dirichlet energy for maps between hyperbolic surfaces is obtained by replacing Euclidean cotangent weights with canonical hyperbolic weights built from the corner angles and edge lengths of the source triangulation. With positive edge weights, the paper's Theorem 2.1 guarantees a unique discrete harmonic map in each homotopy class of the triangulation, it is the unique minimizer of the energy D_c(f)=1/2 ∑ c_ij ℓ_ij^2 over geodesic realizations, and it is an embedding. The weights are positive exactly when the input triangulation is Delaunay, so injectivity is baked into the discretization; and they are the geomet

Load-bearing premise

The theoretical guarantees require all canonical edge weights to be positive, which is equivalent to the input triangulation being Delaunay (the empty-circle condition); for non-Delaunay inputs the algorithm resets negative weights to arbitrary positive numbers, and the paper does not analyze whether uniqueness, embeddedness, or isometry-harmonicity survives that reset.

Editorial extensions

If this is right

  • If the input triangulation is Delaunay (positive weights), the output discrete harmonic map is guaranteed to be an embedding, not merely locally injective.
  • For arbitrary input triangulated surfaces, a discrete-uniformization preprocessing step produces a conformally equivalent hyperbolic Delaunay surface, so the method applies to general genus-2-and-higher meshes.
  • The discrete map inherits the smooth theory's isometry property: when the target is isometric to the source, the computed harmonic map is that isometry.
  • With a sufficiently small time step and an initialization close enough to the solution, gradient descent converges at least linearly; the reported energy and gradient plots exhibit this asymptotic rate.
  • Because the output satisfies the Fuchsian-group constraint globally, it enables remeshing of high-genus surfaces without partitioning them into simply connected pieces.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • A direct test not reported in the paper is to run the algorithm on an isometric pair of hyperbolic surfaces and check whether the output coincides with the known isometry; the theory predicts it should, so this would sharply test the structure-preserving claim.
  • The weight-reset heuristic for non-Delaunay inputs is an unanalyzed gap: the paper does not establish that resetting negative weights to arbitrary positive numbers still yields uniqueness, embeddedness, or isometry-harmonicity, so applications on non-Delaunay meshes cannot rely on the main theorem.
  • Because the energy is proper and its sublevel sets are compact, the gradient-norm stopping criterion gives a rigorous closeness certificate in principle; turning it into a practical a posteriori error bound would be a natural extension.
  • The convergence theorem is local (small step, close initialization), but the experiments suggest a sizable basin of attraction; a step-size schedule or line search that provably preserves convergence while enlarging the basin would be a valuable algorithmic extension.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 4 minor

Summary. The manuscript proposes a structure-preserving numerical method for computing discrete harmonic maps between closed hyperbolic surfaces of genus at least two. The method minimizes a discrete Dirichlet energy (2.3) over geodesic realizations of the 1-skeleton of a triangulated surface in a target hyperbolic surface, using canonical hyperbolic edge weights (2.8). The paper builds on Colin de Verdière's theorem to claim existence, uniqueness and embeddedness of the discrete harmonic map for positive weights, and on Lam's canonical-weight theory to claim that isometries remain harmonic. The numerical scheme lifts the map to the universal cover, enforces equivariance (3.2), and performs Riemannian gradient descent with exponential-map updates (3.9)-(3.10). The authors report experiments on genus-2 and genus-3 surfaces and an application to remeshing.

Significance. If the stated guarantees held, this would be a useful contribution to geometry processing: a global, intrinsic harmonic map for high-genus surfaces with a principled generalization of cotangent weights and a convergence-rate statement. The paper's use of canonical hyperbolic edge weights is a genuine structural idea, and the remeshing application indicates potential practical value. However, the theoretical core is largely imported from prior work, and the paper's own algorithmic modification for non-Delaunay inputs is not covered by the imported theorems. The claims are therefore currently conditional rather than established.

major comments (3)
  1. [§3, negative-edge-weight reset] The theoretical guarantees in Theorem 2.1, §2.3.2 and §3.3 hold only when all canonical edge weights are positive, which is equivalent to the input triangulation being Delaunay. The algorithm states: 'In case the initial surface is not a Delaunay triangulation, the weight of some edges may be negative. To ensure that the final map is an embedding, we will reset the weight of these edges to a prescribed positive number.' This reset changes the energy (2.3) from the canonical energy: the minimizer is no longer the discrete harmonic map associated with the canonical weights, and the harmonicity-of-isometries property (Theorem 2.2) is not preserved. Positivity alone only yields existence, uniqueness and embeddedness for the modified, non-canonical weight system. The paper needs either to restrict the algorithm to Delaunay inputs (e.g., by actually performing edge flips as guaranteed by Theor
  2. [§3.3, Theorem 3.2 applicability] The convergence-rate claim relies on Boumal's theorem for a smooth function on a Riemannian manifold with a retraction. The implemented iteration, however, does not simply run Riemannian gradient descent on D^|V|: it updates one representative of each vertex via the exponential map (3.9) and synchronizes all other representatives via deck transformations (3.10). This coupled iteration is asserted to be equivalent to a Riemannian gradient descent on the quotient manifold or on the equivariant submanifold, but no proof or reference is given for that equivalence, nor is it shown that the synchronized update is a retraction. The statement that 'the special update (3.10) ensures that the map f_t induces a well-defined geodesic realization' is necessary but not sufficient. Without a direct argument, the claimed 'at least linear convergence' from Section 3.3 is not established for the actual al
  3. [§2.4 vs §3] There is an internal inconsistency in the pipeline. Section 2.4 invokes the discrete uniformization theorem to obtain a Delaunay triangulation, and §2.3.2 states that positive canonical weights are equivalent to Delaunay. Yet the algorithm in §3 does not implement edge flips; instead it resets negative weights. If the initial surface is already Delaunay after uniformization, the reset case should not occur; if it is not Delaunay, the algorithm is outside the theoretical framework. The paper should clarify the precise role of discrete uniformization in the algorithm and either guarantee Delaunayness or remove the reset step in favor of flips.
minor comments (4)
  1. [§3, first paragraph] The phrase 'flat discrete hyperbolic metric' is confusing: a hyperbolic metric has curvature -1 and is not flat. Please rephrase.
  2. [§3.2, Eq. (3.4)] The Euclidean-coordinate derivative formula is presented before the Riemannian treatment; since the optimization is Riemannian, it would be cleaner to state from the outset that Euclidean partial derivatives are only used for the initial map and that all descent updates use the Riemannian gradient.
  3. [§3.3, Theorem 3.2] Typo: 'These exists a neighborhood' should be 'There exists a neighborhood'. Also, the theorem statement would benefit from a compactness/hypothesis check: the result is local, and the paper should explicitly identify the neighborhood in the product manifold where the iterates remain.
  4. [§4.2, figures] The text refers to 'Fig. 8c' and 'Fig. 8d' as local regions of the initial and output maps, but the figure captions are not fully consistent with this description. Please verify the figure numbering and labels.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity found: external theorems carry the derivation; the negative-weight reset is a consistency caveat, not a circular step.

full rationale

The paper's derivation chain is self-contained in the sense required here: the central existence/uniqueness/embeddedness statement (Theorem 2.1) is taken from Colin de Verdière [7], the canonical edge-weight formula (2.8) and the isometry-harmonicity property (Theorem 2.2) are imported from published work [20,27] by co-author Lam, and the convergence-rate theorem (Theorem 3.2) is Boumal's external result. No step fits a parameter to the target data and then reports it as a prediction; the energy (2.3), gradient (3.6)-(3.7), and update (3.9) are all explicit and not obtained by inverting the desired output. The self-citations are load-bearing but independent: [27] is a published, parameter-free theorem whose assumptions do not include the present algorithm's output, so under the hard rules it does not raise the circularity score. The one genuine caveat in the manuscript is in Section 3: for non-Delaunay inputs, negative canonical weights are 'reset ... to a prescribed positive number.' This means the algorithm minimizes a modified energy, and the canonical-weight guarantees (e.g., harmonicity of isometries) no longer apply; however, this is a correctness/consistency limitation of the algorithm's scope, not a circular derivation. No equation is reused as its own conclusion.

Assumptions & free parameters 3 free parameters · 8 assumptions · 0 invented entities

The paper's central claim rests on several imported theorems (uniformization, Colin de Verdiere, Poincare, Lam's weight formula, Ricci flow convergence) rather than self-contained derivations. The only genuinely ad hoc inputs are the resetting of negative weights and the unspecified step size, both of which can affect the output but are not analyzed.

free parameters (3)
  • step_size_tau = not specified
    The gradient descent update (3.9) uses a fixed step size tau that is chosen 'small enough' but its value is not reported in the text. It affects convergence and the experimental results.
  • reset_weight_value = not specified
    In Section 3, negative canonical edge weights are reset to a prescribed positive number. This value is not given, and its choice could affect the final map.
  • stopping_threshold_epsilon = not specified
    The stopping criteria (3.11) and (3.12) depend on a prescribed epsilon, which is not specified in the paper.
assumptions (8)
  • standard math Discrete uniformization theorem (Gu-Guo-Luo-Sun-Wu): every closed triangulated genus g>=2 surface with a piecewise Euclidean metric has a unique hyperbolic Delaunay geodesic triangulation in its discrete conformal class.
    Invoked as Theorem 2.3 to justify transforming an arbitrary input surface into a hyperbolic one with Delaunay triangulation.
  • standard math Colin de Verdiere's theorem: for a topological triangulation on a closed non-positive curvature surface with positive edge weights, there exists a unique discrete harmonic map in a given homotopy class, and it is an embedding.
    Used as Theorem 2.1 to establish existence, uniqueness, and embeddedness of the discrete harmonic map.
  • standard math Poincare polygon theorem: a paired geodesic polygon satisfying the cycle condition generates a Fuchsian group and a fundamental domain.
    Used in Section 3.1 to construct the target hyperbolic surface as a quotient of the Poincare disk.
  • domain assumption The canonical hyperbolic edge weights (Lam's formula, equation 2.8) ensure harmonicity of isometries (Theorem 2.2).
    This property is a key motivation for the chosen weights, but it is imported from a prior paper by one of the co-authors and not proven in this manuscript.
  • ad hoc to paper Resetting negative edge weights to an arbitrary positive number preserves the theoretical guarantees of the discrete harmonic map.
    The algorithm does this for non-Delaunay inputs, but the paper does not analyze the effect on uniqueness, embeddedness, or harmonicity of isometries.
  • ad hoc to paper The Riemannian gradient descent with synchronized boundary updates on the universal cover is equivalent to a Riemannian gradient descent on the quotient surface, so Boumal's convergence theorem applies.
    In Section 3.3 the paper assumes this equivalence without a formal proof, which is necessary to transfer the convergence rate result.
  • standard math Hyperbolic Ricci flow converges to the uniformization metric (Chow-Luo).
    Used in Section 3 to equip the input surface with a flat hyperbolic metric before computing edge weights.
  • domain assumption The Hessian of the discrete Dirichlet energy at the discrete harmonic map is positive definite (Kajigaya-Tanaka).
    Relied upon in Section 3.3 to apply the local convergence theorem for Riemannian gradient descent.

how reviews work

0 comments
Cite this review

Pith. "Pith review of A Structure-Preserving Numerical Method for Harmonic Maps Between High-genus Surfaces." pith.science (2026). https://pith.science/paper/2YHNYACN

@misc{pith2026250901256,
  author       = {Pith},
  title        = {Pith review of: A Structure-Preserving Numerical Method for Harmonic Maps Between High-genus Surfaces},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/2YHNYACN}},
  note         = {Machine review of arXiv:2509.01256}
}
read the original abstract

Motivated by geometry processing for surfaces with non-trivial topology, we study discrete harmonic maps between closed surfaces of genus at least two. Harmonic maps provide a natural framework for comparing surfaces by minimizing distortion. Unlike conformal or isometric maps-which may not exist between surfaces with different geometries-harmonic maps always exist within a fixed homotopy class and yield optimal homeomorphisms when the target surface has negative curvature. We develop a structure-preserving algorithm to compute harmonic maps from a triangulated surface to a reference hyperbolic surface. The method minimizes Dirichlet energy over geodesic realizations of the surface graph into the target hyperbolic surface in the homotopy class of a homeomorphism. A central feature of our framework is the use of canonical edge weights derived from the hyperbolic metric, which generalize the classical cotangent weights from the Euclidean setting. These weights preserve injectivity and ensure that isometries remain harmonic in the discrete theory, reflecting their classical behavior.

Figures

Figures reproduced from arXiv: 2509.01256 by the authors.

Figure 1
Figure 1. Two neighboring hyperbolic triangles share a common edge [PITH_FULL_IMAGE:figures/full_fig_p008_1.png] view at source ↗
Figure 2
Figure 2. Construction of a hyperbolic hexagon with paired edges. [PITH_FULL_IMAGE:figures/full_fig_p010_2.png] view at source ↗
Figure 3
Figure 3. The figures show how we compute the descent direction for boundary vertices. [PITH_FULL_IMAGE:figures/full_fig_p013_3.png] view at source ↗
Figures from the paper (11 more)
Figure 4
Figure 4. Figure 4: The figures show how we compute the descent direction for corner vertices. A [PITH_FULL_IMAGE:figures/full_fig_p013_4.png]
Figure 5
Figure 5. Figure 5: A genus-2 example where τ is the step size and ˜f is the final map. As can be seen, the convergence rate rapidly decays from 1 as the iteration approaches the end. This numerical experiment shows that we could expect an at least linear asymptotic convergence rate for s…
Figure 6
Figure 6. Figure 6: A genus-2 example as time passes. 5. An Application to Remeshing. The discrete harmonic map provides an effective way to map a high-genus surface onto a parameter domain globally. Hence, to process the original surface, we could choose to work on the parameter domain f…
Figure 7
Figure 7. Figure 7: A genus-3 example (a) The initial map (b) The output harmonic map (c) A local region of initial map (d) A local region of output map [PITH_FULL_IMAGE:figures/full_fig_p018_7.png]
Figure 8
Figure 8. Figure 8: A genus-3 example [PITH_FULL_IMAGE:figures/full_fig_p018_8.png]
Figure 9
Figure 9. Figure 9: A genus-3 example (a) Descent of energy for a larger step (b) Descent of gradient for a larger step [PITH_FULL_IMAGE:figures/full_fig_p019_9.png]
Figure 10
Figure 10. Figure 10: A genus-3 example [PITH_FULL_IMAGE:figures/full_fig_p019_10.png]
Figure 11
Figure 11. Figure 11: The given surface to be remeshed first notice that ˜f(S˜) does not necessarily cover P. But since P is compact, we could find a discrete subset {γ1, · · · , γk} ⊂ Γ such that (5.1) P ⊂ [ k i=1 γi( ˜f(S˜)) Recall the fact that for any γ ∈ Γ, we have πV (i) = πV (γ(i)) …
Figure 12
Figure 12. Figure 12: The mesh structures on target surfaces [PITH_FULL_IMAGE:figures/full_fig_p021_12.png]
Figure 13
Figure 13. Figure 13: The harmonic maps from the input surface to the target ones [PITH_FULL_IMAGE:figures/full_fig_p021_13.png]
Figure 14
Figure 14. Figure 14: The two remeshed surfaces (2003), pp. 233–256. [9] M. Desbrun, M. Meyer, and P. Alliez, Intrinsic parameterizations of surface meshes, in Computer graphics forum, vol. 21, Wiley Online Library, 2002, pp. 209–218. [10] R. J. Duffin, Distributed and lumped networks, J. …

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Morphing Graphs on Hyperbolic Surfaces

    math.GT 2026-07 conditional novelty 8.0 of 10

    First algorithm for morphing geodesic graph embeddings on hyperbolic surfaces, built on a generalized Tutte embedding theorem and hyperbolic mean value coordinates.

Reference graph

Works this paper leans on

48 extracted references · 43 canonical work pages · cited by 1 Pith paper

  1. [1]

    Angenent, S

    S. Angenent, S. Haker, A. Tannenbaum, and R. Kikinis , Conformal geometry and brain flattening, in Medical Image Computing and Computer-Assisted Intervention–MICCAI’99: Second International Conference, Cambridge, UK, September 19-22, 1999. Proceedings 2, Springer, 1999, pp. 271–278

    work page 1999

  2. [2]

    Bajcsy and S

    R. Bajcsy and S. Kovaˇciˇc, Multiresolution elastic matching , Computer vision, graphics, and image processing, 46 (1989), pp. 1–21

    work page 1989

  3. [3]

    M. F. Beg, M. I. Miller, A. Trouv ´e, and L. Younes , Computing large deformation metric mappings via geodesic flows of diffeomorphisms , International journal of computer vision, 61 (2005), pp. 139–157

    work page 2005

  4. [4]

    A. I. Bobenko, U. Pinkall, and B. A. Springborn , Discrete conformal maps and ideal hyperbolic polyhedra, Geom. Topol., 19 (2015), pp. 2155–2215, https://doi.org/10.2140/gt. 2015.19.2155, https://doi.org/10.2140/gt.2015.19.2155

    doi:10.2140/gt 2015

  5. [5]

    Boumal, An introduction to optimization on smooth manifolds, Cambridge University Press, 2023

    N. Boumal, An introduction to optimization on smooth manifolds, Cambridge University Press, 2023

    2023

  6. [6]

    Chow and F

    B. Chow and F. Luo, Combinatorial ricci flows on surfaces , Journal of Differential Geometry, 63 (2003), pp. 97–129

    work page 2003

  7. [7]

    Colin de Verdi`ere, Comment rendre g´ eod´ esique une triangulation d’une surface?, Enseign

    Y. Colin de Verdi`ere, Comment rendre g´ eod´ esique une triangulation d’une surface?, Enseign. Math. (2), 37 (1991), pp. 201–212

    work page 1991

  8. [8]

    C. R. Collins and K. Stephenson , A circle packing algorithm , Computational Geometry, 25 22 ZHIPENG ZHU, W AI YEUNG LAM, AND LOK MING LUI Fig. 14: The two remeshed surfaces (2003), pp. 233–256

    work page 2003

Show all 48 references
  1. [9]

    Desbrun, M

    M. Desbrun, M. Meyer, and P. Alliez , Intrinsic parameterizations of surface meshes , in Computer graphics forum, vol. 21, Wiley Online Library, 2002, pp. 209–218

    2002

  2. [10]

    R. J. Duffin, Distributed and lumped networks , J. Math. Mech., 8 (1959), pp. 793–826, https: //doi.org/10.1512/iumj.1959.8.58051, https://doi.org/10.1512/iumj.1959.8.58051

    1959 doi

  3. [11]

    M. Eck, T. DeRose, T. Duchamp, H. Hoppe, M. Lounsbery, and W. Stuetzle , Multires- olution analysis of arbitrary meshes , in Proceedings of the 22nd annual conference on Computer graphics and interactive techniques, 1995, pp. 173–182

    1995

  4. [12]

    Erickson and P

    J. Erickson and P. Lin , A Toroidal Maxwell-Cremona-Delaunay Correspondence , in 36th International Symposium on Computational Geometry (SoCG 2020), S. Cabello and D. Z. Chen, eds., vol. 164 of Leibniz International Proceedings in Informatics (LIPIcs), Dagstuhl, Germany, 2020,...

    2020 doi

  5. [13]

    M. S. Floater , Parametrization and smooth approximation of surface triangulations , Com- puter aided geometric design, 14 (1997), pp. 231–250

    1997

  6. [14]

    Gaster, B

    J. Gaster, B. Loustau, and L. Monsaingeon , Computing discrete equivariant har- monic maps , 2018, https://arxiv.org/abs/1810.11932, https://arxiv.org/abs/1810.11932. arXiv:1810.11932

    2018 arXiv

  7. [15]

    S. J. Gortler, C. Gotsman, and D. Thurston , Discrete one-forms on meshes and ap- plications to 3d mesh parameterization , Computer Aided Geometric Design, 23 (2006), pp. 83–112

    2006

  8. [16]

    X. Gu, R. Guo, F. Luo, J. Sun, and T. Wu, A discrete uniformization theorem for polyhedral surfaces II , J. Differential Geom., 109 (2018), pp. 431–466, https://doi.org/10.4310/jdg/ 1531188190, https://doi.org/10.4310/jdg/1531188190

    2018

  9. [17]

    Gu and S.-T

    X. Gu and S.-T. Yau , Global conformal surface parameterization, in Proceedings of the 2003 Eurographics/ACM SIGGRAPH symposium on Geometry processing, 2003, pp. 127–137

    2003

  10. [18]

    Haker, L

    S. Haker, L. Zhu, A. Tannenbaum, and S. Angenent , Optimal mass transport for registra- tion and warping , International Journal of computer vision, 60 (2004), pp. 225–240

    2004

  11. [19]

    Isenburg and P

    M. Isenburg and P. Lindstrom , Streaming meshes , in VIS 05. IEEE Visualization, 2005., IEEE, 2005, pp. 231–238

    2005

  12. [20]

    Izmestiev and W

    I. Izmestiev and W. Y. Lam , Discrete laplacians – spherical and hyperbolic , J. Lond. Math. Soc. (2), (2025), https://doi.org/10.1112/jlms.70235

    2025 doi

  13. [21]

    M. Jin, J. Kim, F. Luo, and X. Gu , Discrete surface ricci flow , IEEE Transactions on Visu- alization and Computer Graphics, 14 (2008), pp. 1030–1043

    2008

  14. [22]

    Jost , Harmonic maps between surfaces , vol

    J. Jost , Harmonic maps between surfaces , vol. 1062 of Lecture Notes in Mathematics, Springer-Verlag, Berlin, 1984, https://doi.org/10.1007/BFb0100160, https://doi.org/10. 1007/BFb0100160. NUMERICAL METHOD FOR HARMONIC MAPS BETWEEN HYPERBOLIC SURF ACES23

    1984 doi

  15. [23]

    J. Jost , Riemannian geometry and geometric analysis , Universitext, Springer, Cham, seventh ed., 2017, https://doi.org/10.1007/978-3-319-61860-9, https://doi.org/10.1007/ 978-3-319-61860-9

    2017 doi

  16. [24]

    Kajigaya and R

    T. Kajigaya and R. Tanaka , Uniformizing surfaces via discrete harmonic maps , Ann. H. Lebesgue, 4 (2021), pp. 1767–1807, https://doi.org/10.5802/ahl.116, https://doi.org/10. 5802/ahl.116

    2021 doi

  17. [25]

    Kotani and T

    M. Kotani and T. Sunada, Standard realizations of crystal lattices via harmonic maps, Trans. Amer. Math. Soc., 353 (2001), pp. 1–20, https://doi.org/10.1090/S0002-9947-00-02632-5, https://doi.org/10.1090/S0002-9947-00-02632-5

    2001 doi

  18. [26]

    W. Y. Lam, Delaunay decompositions minimizing energy of weighted toroidal graphs , Discrete Comput. Geom., (2024), https://doi.org/10.1007/s00454-024-00699-x

    2024 doi

  19. [27]

    W. Y. Lam, Discrete harmonic maps between hyperbolic surfaces, Calc. Var. Partial Differential Equations, (2025), https://arxiv.org/abs/2405.02205. To appear

    2025 arXiv

  20. [28]

    R. Li, T. Tang, and P. Zhang , Moving mesh methods in multiple dimensions based on har- monic maps , Journal of Computational Physics, 170 (2001), pp. 562–588

    2001

  21. [29]

    L. M. Lui, K. C. Lam, T. W. Wong, and X. Gu , Texture map and video compression using beltrami representation, SIAM Journal on Imaging Sciences, 6 (2013), pp. 1880–1902

    2013

  22. [30]

    L. M. Lui and C. Wen , Geometric registration of high-genus surfaces , SIAM Journal on Imaging Sciences, 7 (2014), pp. 337–365

    2014

  23. [31]

    L. M. Lui, T. W. Wong, W. Zeng, X. Gu, P. M. Thompson, T. F. Chan, and S. T. Yau, Detection of shape deformities using yamabe flow and beltrami coefficients , Inverse Problems and Imaging, 4 (2010), pp. 311–333

    2010

  24. [32]

    Luo , Combinatorial Yamabe flow on surfaces , Commun

    F. Luo , Combinatorial Yamabe flow on surfaces , Commun. Contemp. Math., 6 (2004), pp. 765–780, https://doi.org/10.1142/S0219199704001501, https://doi.org/10. 1142/S0219199704001501

    2004 doi

  25. [33]

    Marchandise, C

    E. Marchandise, C. C. de Wiart, W. Vos, C. Geuzaine, and J.-F. Remacle , High- quality surface remeshing using harmonic maps—part ii: Surfaces with high genus and of large aspect ratio, International Journal for Numerical Methods in Engineering, 86 (2011), pp. 1303–1321

    2011

  26. [34]

    D. B. Mumford and J. Shah , Optimal approximations by piecewise smooth functions and as- sociated variational problems, Communications on pure and applied mathematics, (1989)

    1989

  27. [35]

    T. Ng, X. Gu, and L. Lui , Teichm¨ uller extremal map of multiply-connected domains using beltrami holomorphic flow , Journal of Scientific Computing, 60 (2014), pp. 249–275

    2014

  28. [36]

    Persson and G

    P.-O. Persson and G. Strang , A simple mesh generator in matlab , SIAM review, 46 (2004), pp. 329–345

    2004

  29. [37]

    Pinkall and K

    U. Pinkall and K. Polthier , Computing discrete minimal surfaces and their conjugates , Experimental mathematics, 2 (1993), pp. 15–36

    1993

  30. [38]

    Remacle, C

    J.-F. Remacle, C. Geuzaine, G. Compere, and E. Marchandise , High-quality surface remeshing using harmonic maps , International Journal for Numerical Methods in Engi- neering, 83 (2010), pp. 403–425

    2010

  31. [39]

    T. W. Sederberg and F. Chen , Implicitization using moving curves and surfaces , in Pro- ceedings of the 22nd annual conference on Computer graphics and interactive techniques, 1995, pp. 301–308

    1995

  32. [40]

    W. T. Tutte, How to draw a graph, Proceedings of the London Mathematical Society, 3 (1963), pp. 743–767

    1963

  33. [41]

    V aillant, M

    M. V aillant, M. I. Miller, L. Younes, and A. Trouv ´e, Statistics on diffeomorphisms via tangent space representations, NeuroImage, 23 (2004), pp. S161–S169

    2004

  34. [42]

    W ang, Y

    S. W ang, Y. W ang, M. Jin, X. D. Gu, and D. Samaras, Conformal geometry and its appli- cations on 3d shape matching, recognition, and stitching , IEEE Transactions on Pattern Analysis and Machine Intelligence, 29 (2007), pp. 1209–1220

    2007

  35. [43]

    G. Xu, B. Mourrain, R. Duvigneau, and A. Galligo , Parameterization of computational domain in isogeometric analysis: methods and comparison , Computer Methods in Applied Mechanics and Engineering, 200 (2011), pp. 2021–2031

    2011

  36. [44]

    W. Zeng, L. M. Lui, F. Luo, T. F.-C. Chan, S.-T. Yau, and D. X. Gu , Computing quasi- conformal maps using an auxiliary metric and discrete curvature flow , Numerische Math- ematik, 121 (2012), pp. 671–703

    2012

  37. [45]

    Zhang and M

    D. Zhang and M. Hebert , Harmonic maps and their applications in surface matching , in Proceedings. 1999 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (Cat. No PR00149), vol. 2, IEEE, 1999, pp. 524–530

    1999

  38. [46]

    Zhang, R

    M. Zhang, R. Liao, A. V. Dalca, E. A. Turk, J. Luo, P. E. Grant, and P. Golland , Frequency diffeomorphisms for efficient image registration , in Information Processing in Medical Imaging: 25th International Conference, IPMI 2017, Boone, NC, USA, June 25-30, 24 ZHIPENG ZHU, W ...

    2017

  39. [47]

    Zhou and A

    Q. Zhou and A. Jacobson, Thingi10k: A dataset of 10,000 3d-printing models , arXiv preprint arXiv:1605.04797, (2016)

    2016 arXiv

  40. [48]

    Z. Zhu, G. P. Choi, and L. M. Lui , Parallelizable global quasi-conformal parameterization of multiply connected surfaces via partial welding , SIAM Journal on Imaging Sciences, 15 (2022), pp. 1765–1807

    2022

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.