Pith. sign in

REVIEW 2 major objections 5 minor 1 cited by

A variational method for generating $n$-cross fields using higher-order $Q$-tensors

T0 review · 2 major / 5 minor · reviewed 2026-08-14 · deepseek-v4-flash

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

Pith's one-line read This paper establishes that an n-cross field is exactly a symmetric fourth-order tensor with $Q^2=Q$ and block traces, and uses Ginzburg-Landau relaxation to generate such fields on Lipschitz domains.

desk verdict A useful variational framework for n-cross fields with a real proof gap: the central representation theorem's algebraic step is unjustified, and the abstract overclaims what is proven. read the letter →

arxiv 1909.00922 v3 pith:UTS5PI3N submitted 2019-09-03 math.AP cs.CG

Dmitry Golovaty , Jose Alberto Montero , Daniel Spirn This is my paper · ORCID
classification math.APcs.CG MSC 35Q5615A69
keywords n-crossfieldsfourth-orderQ-tensorsGinzburg-LandaurelaxationorthogonalprojectionmatricestensorvarietiesLipschitzdomainscross-fieldsingularitiesframe
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

The paper aims to show that generating an n-cross field—n mutually orthogonal line directions assigned to each point, with the symmetries of the cube's rotation group—can be treated as a problem in the calculus of variations. Its central claim is that n-cross fields are exactly the symmetric fourth-order tensors $Q$ satisfying $Q^2=Q$ and block-trace conditions $\operatorname{tr} Q_{ij}=\delta_{ij}$, because every such tensor splits as a sum of $n$ rank-one orthogonal projections with pairwise orthogonal images. That equivalence turns field generation into a Ginzburg-Landau relaxation: minimize a Dirichlet energy plus a penalty measuring $Q^2-Q$, with boundary terms that force one line of the field to be normal to the boundary. The method thus yields a dimension-independent variational route to cross fields and places their singularities inside a classical PDE framework.

What carries the argument

The load-bearing object is the fourth-order Q-tensor read as an $n\times n$ block matrix of $n\times n$ blocks, with permutation symmetry, idempotence $Q^2=Q$, and block traces $\operatorname{tr} Q_{ij}=\delta_{ij}$ cutting out $M^n_{\mathrm{cross}}$ as a polynomial variety. The proof of the representation theorem passes through the isomorphism $\Phi_0$ that sends an $n^2\times n^2$ matrix to a linear map on $n\times n$ matrices; a spectral decomposition of $Q$ under that map produces symmetric matrices $Q_j$, and the critical step is proving that these $Q_j$ commute. That step uses an identity expressing $Q^4$ in terms of $\langle Q_jQ_k,A\rangle Q_jQ_k$, together with permutation invariance of $Q^2$ and $Q^4$; testing the identity against antisymmetric matrices forces the commutators $[Q_j,Q_k]$ to vanish. Commuting symmetric matrices then share an eigenframe, which yields the projections $P_j$. The same machinery identifies the odeco variety of orthogonally decomposable tensors with the same class of sums, up to eigenvalues.

What would settle it

Take $n=2$, choose two orthonormal symmetric matrices that do not commute, and compute both sides of the asserted identity $Q^4_L(A)=\sum_{j,k}\lambda_j^2\lambda_k^2\langle Q_jQ_k,A\rangle Q_jQ_k$ for an antisymmetric test matrix $A$; if the two sides differ, the commutativity step has no valid basis. A direct way to run the check is to form $Q^4$ by composing the block formula for $Q^2$ rather than using the asserted sum, and compare the antisymmetric part.

Watch

Extended reading notes

Core claim

The central discovery, stated as Theorem 4.2, is that the algebraic conditions defining the tensor set $M^n_{\mathrm{cross}}$ are sufficient as well as necessary: any symmetric fourth-order tensor $Q$ with $Q^2=Q$ and $\operatorname{tr} Q_{ij}=\delta_{ij}$ can be written as $Q=\sum_{j=1}^n P_j\otimes P_j$ for rank-one orthogonal projections $P_j$ whose images are pairwise perpendicular. Consequently the abstract object “n-cross” and the concrete object “idempotent symmetric tensor with the right block traces” are the same set. A corollary gives a recovery rule: since the blocks $Q_{ij}$ share a common eigenframe, the n-cross is read off by diagonalizing any single block. The paper then relaxes the idempotence constraint by adding the potential $W(Q)=|Q^2-Q|^2$ to a Dirichlet energy, and it encodes boundary alignment through three equivalent conditions: the boundary normal belongs to the frame, its associated projection commutes with $Q$, or each block satisfies $Q_{ij}\nu=\nu_i\nu_j\nu$.

Load-bearing premise

The proof that the spectral pieces commute rests on a block-matrix identity for $Q^4$ that is asserted right after a calculation; if that identity is not valid as written, the theorem's conclusion is not established.

Editorial extensions

If this is right

  • Because any block $Q_{ij}$ of an admissible tensor already determines the whole n-cross through its eigenframe, field recovery is a byproduct of the relaxation rather than a separate step.
  • Minimizing the relaxed energy with $|Q^2-Q|^2$ as the penalty term yields, as $\varepsilon$ tends to zero, maps that are n-cross fields almost everywhere, with singularities located where the tensor fails to be a projection.
  • The boundary conditions admit three equivalent formulations, so alignment with a prescribed normal can be imposed either as a hard constraint or as a weak-anchoring penalty, both fitting into standard Ginzburg-Landau theory.
  • In two dimensions the construction reproduces the known quartic energy for degree-$1/4$ vortices, linking the higher-dimensional method to an established theory.
  • Numerical gradient-flow solutions reproduce known 3-cross configurations, including eight vortices on a spherical shell and four disclination lines in a torus with a hole, showing that the topological singular structure is captured.

Reading between the lines

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

  • Inference: the same polynomial description may give an explicit coordinate chart for the quotient space $SO(n)/O_n$ in any dimension, which would make topological invariants of cross-field singularities computable without reconstructing frames.
  • Inference: an explicit nearest-n-cross projection could be built from the eigenframe of any block $Q_{ij}$, and if it is Lipschitz it would enable a fast projection-and-solve evolution whose limiting fields could be compared field-by-field with the gradient-flow ones.
  • Inference: in four dimensions the boundary condition leaves a circle's worth of frames at each boundary point, so the same method should produce one-dimensional singular strata; computing 4-cross fields numerically would separate generic topological effects from special three-dimensional ones.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

2 major / 5 minor

Summary. The manuscript proposes a variational framework for generating n-cross fields. It represents an n-cross by a fourth-order tensor Q = Σ_j P_j⊗P_j and defines the set M^n_cross as symmetric tensors with block traces tr(Q_ij)=δ_ij and Q^2=Q. Theorem 4.2 asserts the converse: every element of M^n_cross is of that form. On this basis, the authors introduce two Ginzburg-Landau-type energies with bulk and boundary penalization, derive explicit gradient-flow PDEs for n=2,3, and present numerical examples in 3D (notched cube, spherical shell, ball, torus). The paper explicitly defers the Γ-convergence and limiting analysis to future work, while the abstract states that the relaxation 'reliably generates' an n-cross field.

Significance. If Theorem 4.2 holds, the tensor characterization gives a coordinate-invariant description of n-cross fields in arbitrary dimensions, and the Ginzburg-Landau relaxation is a natural PDE-based construction with a new selection principle. The boundary-anchoring condition via commutators in Proposition 3.1 and the explicit evolution system in Appendix B are useful contributions. The numerical examples reproduce structures from prior work, and the paper is honest about several open questions, including the Γ-limit and the relationship to MBO-type schemes. However, the central representation theorem is currently not established by the proof as written, so the significance of the framework is conditional on repairing that proof.

major comments (2)
  1. [Appendix A, proof of Theorem 4.2] The identity displayed after the block-matrix computation in the proof of Theorem 4.2, Q^4_L(A)=Σ_j λ_j^4⟨Q_j,A⟩Q_j = Σ_{i,j} λ_i^2λ_j^2⟨Q_iQ_j,A⟩Q_iQ_j, is not justified. Applying Φ0 to the block matrix with (i,j)-block (Q_iQ_j)_{ab}Q_iQ_j gives, by Eq. (9.6), the linear map L_{Q_iQ_j}(A)=Q_iQ_j A Q_jQ_i, not the rank-one operator A↦⟨Q_iQ_j,A⟩Q_iQ_j. The subsequent decomposition into symmetric and antisymmetric parts, displayed leading to Eq. (9.13), and the conclusion that the Q_j commute rely exactly on this incorrect identity. Therefore the proof of Theorem 4.2, and with it Corollary 4.3, Proposition 3.1, and the interpretation of the numerical minimizers as elements of M^n_cross, is not established as written. The theorem may be salvageable, but a corrected derivation of the commutativity step is required.
  2. [Abstract; §4.2; §8] The abstract claims that 'one can reliably generate an n-cross field' by the Ginzburg-Landau relaxation, but no convergence or Γ-convergence theorem is proved. Section 4.2 states that the analysis of the variational problem is left to a follow-up paper, and Section 8 lists the Γ-limit as an open problem. The numerical experiments, while suggestive, are qualitative visual comparisons rather than evidence of rigorous convergence. The claim in the abstract should either be supported by a theorem or weakened to describe a proposed method with numerical support.
minor comments (5)
  1. [Abstract] The sentence 'It was recently by other authors that 3-cross fields...' is missing a verb and should read 'It was recently shown by other authors that 3-cross fields...'.
  2. [Theorem 4.2] In the statement of Theorem 4.2, 'for all j,k = 1,...' is missing its upper limit and should read 'for all j,k = 1,...,n'.
  3. [Lemma 6.1 proof] In the proof of Lemma 6.1, the line 'For Q23 we use Q2311 = Q1123, Q2312 = Q1123' contains a typo; the second identity should presumably be Q2312 = Q1223.
  4. [§7.4] The final sentence of Section 7.4 refers to 'cross-sections of the 3-cross field in the sphere', but the domain in that subsection is a toroid with a cylindrical hole; this should be reworded to avoid confusion.
  5. [Proposition 3.1] In Proposition 3.1 and its proof, the symbol P is used both for a rank-one projection matrix and for the block tensor P⊗P; the distinction should be made explicit, since the commutator condition [Q,P]=0 depends on which object is meant.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the tensor characterization is proved from the algebraic constraints, and the numerical comparisons are qualitative checks rather than fitted predictions.

full rationale

The derivation is self-contained. The canonical tensor Q = Σ P_j ⊗ P_j is built from an n-cross, and Lemmas 2.1–2.4 establish that such Q is symmetric, idempotent, and has block traces tr(Q_ij) = δ_ij. Theorem 4.2 is the converse: starting from a symmetric tensor with Q^2 = Q and tr(Q_ij) = δ_ij, Appendix A uses the spectral decomposition of Φ0(Q), the permutation equivariance Tσ, and the trace conditions to produce commuting rank-one projections. The algebraic constraints are hypotheses and the projection decomposition is the conclusion; neither is assumed to prove the other, so this is a genuine characterization rather than a self-definitional reduction. The Ginzburg-Landau energy is intentionally designed with M^n_cross as the zero set of the penalized potential, so the ε → 0 limit being an n-cross field is a consequence of Theorem 4.2 and not a fitted input renamed as a prediction. Comparisons with [25] and [33] are qualitative visual checks, not benchmarks used to fit model parameters. The self-citations [1] and [14] are background analogies (weak-anchoring singular-set migration and the Landau–de Gennes penalty) and are not load-bearing for the main theorem. A possible algebraic gap in Appendix A’s commutativity argument (the operator identity following the block computation may not follow as written) is a correctness concern, not a circular reduction, and therefore does not affect the circularity score.

Assumptions & free parameters 2 free parameters · 4 assumptions · 0 invented entities

No new particles or forces. The free parameters are numerical regularization parameters, not fitted constants. The main unstated input is the well-posedness of the gradient flow and the correctness of the disputed algebraic identity in Appendix A.

free parameters (2)
  • epsilon (Ginzburg-Landau relaxation parameter) = about 0.1 * domain size; 0.02 in Section 7.2
    Chosen by hand for simulations; controls how strongly Q is forced to satisfy Q^2 = Q. The central mathematical results are asymptotic in epsilon, but numerical outputs depend on it.
  • delta_epsilon (boundary relaxation parameter) = same values as epsilon in the examples
    Chosen by hand; controls boundary penalty strength; not analyzed.
assumptions (4)
  • standard math Spectral theorem for symmetric matrices on finite-dimensional spaces
    Used in Appendix A to decompose Q_L into eigenprojections.
  • standard math Poincare-Hopf theorem for tangent vector fields on spheres
    Used in Section 3 to show obstruction to smooth boundary-aligned cross fields on a ball.
  • domain assumption Existence and sufficient regularity of gradient flow solutions for the COMSOL-solved PDE system on Lipschitz domains
    The paper solves d_t q = -grad E numerically but provides no well-posedness or convergence analysis; this is acknowledged in Sections 4.2 and 8 as future work.
  • domain assumption The boundary condition that a line of the cross field is normal to the boundary can be encoded by equation (3.2)
    This is the modeling assumption for mesh-aligned cross fields; Proposition 3.1 shows it is equivalent to [Q,P] = 0 for Q in M^n_cross, but the choice of this boundary condition is an application-driven modeling decision.

how reviews work

0 comments
Cite this review

Pith. "Pith review of A variational method for generating $n$-cross fields using higher-order $Q$-tensors." pith.science (2026). https://pith.science/paper/UTS5PI3N

@misc{pith2026190900922,
  author       = {Pith},
  title        = {Pith review of: A variational method for generating $n$-cross fields using higher-order $Q$-tensors},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/UTS5PI3N}},
  note         = {Machine review of arXiv:1909.00922}
}
abstract

An $n$-cross field is a locally-defined orthogonal coordinate system invariant with respect to the cubic symmetry group. Cross fields are finding wide-spread use in mesh generation, computer graphics, and materials science among many applications. It was recently by other authors that $3$-cross fields can be embedded into the set of symmetric $4$th-order tensors. Another concurrent work further develops a relaxation of this tensor field via a certain set of varieties. In this paper, we consider the problem of generating an arbitrary $n$-cross field using a fourth-order $Q$-tensor theory that is constructed out of tensored projection matrices. We establish that by a Ginzburg-Landau relaxation towards a global projection, one can reliably generate an $n$-cross field on arbitrary Lipschitz domains. Our work provides a rigorous approach that offers several new results including porting the tensor framework to arbitrary dimensions, providing a new relaxation method that embeds the problem into a global steepest descent, and offering a relaxation scheme for aligning the cross field with the boundary. Our approach is designed to fit within the classical Ginzburg-Landau PDE theory, offering a concrete road map for the future careful study of singularities of energy minimizers.

Figures

Figures reproduced from arXiv: 1909.00922 by the authors.

Figure 1
Figure 1. Solution of the Ginzburg-Landau PDE subject to the natural boundary conditions in a cube-shaped domain with a cylindrical notch: A disclination in the 3-cross field via level surface plot for 1 ε 2 |Q 2 − Q| 2 (left); Streamlines of the 3-cross solution field with colors representing three different directions followed along the cross field (right) [PITH_FULL_IMAGE:figures/full_fig_p005_1.png] view at source ↗
Figure 2
Figure 2. Solution of the Ginzburg-Landau PDE subject to the natural boundary conditions in a cube-shaped domain with a cylindrical notch: A horizontal cross￾section of the 3-cross solution field at the level intersecting the notch [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Solution of the Ginzburg-Landau PDE subject to the natural boundary conditions in a bone-shaped domain. Note that the singular set (marked in red) does not exhibit twisting (cf. [25]). 2. n-crosses via higher order Q-tensors In this section we rigorously develop a tensor representation of n-cross fields. In the following discussion we distinguish a vector as a, a square matrix as A, and a tensor as A. Let A, B be tw… view at source ↗
Figures from the paper (10 more)
Figure 4
Figure 4. Figure 4: Cross-sections of the 3-cross field (3.3)-(3.5) along xy- and yz-planes, respectively (top row). The 3-cross field (3.3)-(3.5) on the surface of the sphere (bottom row). The singular point is located at the south pole of the sphere. The vectors a 1 , a 2 , and a 3 are …
Figure 5
Figure 5. Figure 5: The ”escaped” 3-cross field (3.6)-(3.8) in the cylinder. The vectors a 1 , a 2 , and a 3 are marked in different colors to aid visualization. The vector field a 3 normal to the boundary, avoids singularity by escaping into the third dimension, i.e., reorienting along t…
Figure 6
Figure 6. Figure 6: Vortices on a thin spherical shell. The figure shows the contour plot of the potential W(Q). the domain size. Note that there is a relationship between ε and δε that determines whether the topological defects of minmimizers of (6.2) lie on the boundary or the interior …
Figure 7
Figure 7. Figure 7: The 3-cross distribution on the shaded one eighth of the spherical shell in [PITH_FULL_IMAGE:figures/full_fig_p025_7.png]
Figure 8
Figure 8. Figure 8: Singularities in a ball. The contour plot of the potential W(Q) indicates that there are eight surface vortices connected by eight disclination lines. The lengths of the frame vectors inside the disclination cores are scaled to make the intersections between the discli…
Figure 9
Figure 9. Figure 9: Cross-section along the xy-plane of the 3-cross field inside a ball. see in the right inset in [PITH_FULL_IMAGE:figures/full_fig_p026_9.png]
Figure 11
Figure 11. Figure 11: 7.5. Domains with complex geometries. The last set of examples ( [PITH_FULL_IMAGE:figures/full_fig_p026_11.png]
Figure 10
Figure 10. Figure 10: Disclinations on the toroidal domain with a cylindrical hole. The contour plot of the potential W(Q) indicates that there are four disclination lines in the part of the domain away from the hole. generic singular sets, as have been seen numerically and analytically. T…
Figure 11
Figure 11. Figure 11: Cross-sections of the 3-cross field inside the toroidal domain with a cylindrical hole along xy-, xz-, and yz-planes, respectively. Finally, it is of interest to see if there is an explicit way to generate the nearest n-cross field from any fourth order symmetric tens…
Figure 12
Figure 12. Figure 12: Disclinations in the domains with complex geometries. The right column shows the contour plot of the potential W(Q) associated the domain shown in the left column. 9.1. Notation. In this section L(F) denotes the set of linear maps from the vector space F to itself. Th…

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. Algebraic Representations for Volumetric Frame Fields

    cs.GR 2019-08 conditional novelty 8.0 of 10

    Octahedral frame space embeds isometrically in R9, giving closed-form geodesics and SDP projection, and a new odeco frame type handles singular curves.

Reference graph

Works this paper leans on

33 extracted references · 32 canonical work pages · cited by 1 Pith paper

  1. [7]

    Chemin, F

    A. Chemin, F. Henrotte, J.-F. Remacle, and J. V. Schaftingen. Representing Three-Dimensional Cross Fields Using Fourth Order Tensors, pages 89–108. Springer International Publishing, Cham, 2019

    work page 2019

  2. [25]

    Algebraic Representations for Volumetric Frame Fields

    D. Palmer, D. Bommes, and J. Solomon. Algebraic representations for volumetric frame fields. arXiv preprint arXiv:1908.05411, 2019

    work page Pith review arXiv 1908

  3. [1]

    Thin Film Liquid Crystals with Oblique Anchoring and Boojums

    S. Alama, L. Bronsard, and D. Golovaty. Thin film liquid crystals with oblique anchoring and boojums. arXiv preprint arXiv:1907.04757, 2019

    work page Pith review arXiv 1907

  4. [2]

    Bauman, D

    P. Bauman, D. Phillips, and C. Wang. Higher dimensional Ginzburg-Landau equations under weak anchoring boundary conditions. Journal of Functional Analysis , 276(2):447 – 495, 2019

    work page 2019

  5. [3]

    Beaufort, J

    P.-A. Beaufort, J. Lambrechts, F. Henrotte, C. Geuzaine, and J.-F. Remacle. Computing cross fields a pde approach based on the ginzburg-landau theory. Procedia Engineering, 203:219 – 231, 2017. 26th International Meshing Roundtable, IMR26, 18-21 September 2017, Barcelona, Spain

    work page 2017

  6. [4]

    Bernard, J.-F

    P.-E. Bernard, J.-F. Remacle, N. Kowalski, and C. Geuzaine. Hex-dominant meshing approach based on frame field smoothness. Procedia Engineering, 82:175 – 186, 2014. 23rd International Meshing Roundtable (IMR23)

    work page 2014

  7. [5]

    B. A. Bilby, R. Bullough, E. Smith, and J. M. Whittaker. Continuous distributions of dislocations: a new application of the methods of non-riemannian geometry. Proceedings of the Royal Society of London. Series A. Mathematical and Physical Sciences , 231(1185):263–273, 1955

    work page 1955

  8. [6]

    Bommes, H

    D. Bommes, H. Zimmer, and L. Kobbelt. Mixed-integer quadrangulation. ACM Trans. Graph. , 28(3):77:1– 77:10, July 2009

    work page 2009

Show all 33 references
  1. [8]

    COMSOL Multiphysics R© v. 5.3. http://www.comsol.com/. COMSOL AB, Stockholm, Sweden

  2. [9]

    Corman and K

    E. Corman and K. Crane. Symmetric moving frames. ACM Trans. Graph., 38(4):87:1–87:16, July 2019

    2019

  3. [10]

    Epstein and R

    M. Epstein and R. Segev. Geometric aspects of singular dislocations. Mathematics and Mechanics of Solids , 19(4):337–349, 2014

    2014

  4. [11]

    X. Fang, W. Xu, H. Bao, and J. Huang. All-hex meshing using closed-form induced polycube. ACM Trans. Graph., 35(4):124:1–124:9, July 2016

    2016

  5. [12]

    H. J. Fogg, L. Sun, J. E. Makem, C. G. Armstrong, and T. T. Robinson. Singularities in structured meshes and cross-fields. Computer-Aided Design, 105:11 – 25, 2018. A VARIATIONAL METHOD FOR GENERATING n-CROSS FIELDS USING HIGHER-ORDER Q-TENSORS 39

    2018

  6. [13]

    X. Gao, W. Jakob, M. Tarini, and D. Panozzo. Robust hex-dominant mesh generation using field-guided polyhedral agglomeration. ACM Trans. Graph., 36(4):114:1–114:13, July 2017

    2017

  7. [14]

    Golovaty and J

    D. Golovaty and J. A. Montero. On minimizers of a Landau–de Gennes energy functional on planar domains. Arch. Ration. Mech. Anal., 213(2):447–490, 2014

    2014

  8. [15]

    Huang, Y

    J. Huang, Y. Tong, H. Wei, and H. Bao. Boundary aligned smooth 3d cross-frame field. ACM Trans. Graph., 30(6):143:1–143:8, Dec. 2011

    2011

  9. [16]

    Kleman and O

    M. Kleman and O. D. Lavrentovich. Topological point defects in nematic liquid crystals. Philosophical Maga- zine, 86(25-26):4117–4137, 2006

    2006

  10. [17]

    Kn¨ oppel, K

    F. Kn¨ oppel, K. Crane, U. Pinkall, and P. Schr¨ oder. Globally optimal direction fields. ACM Trans. Graph. , 32(4), July 2013

    2013

  11. [18]

    Kowalski, F

    N. Kowalski, F. Ledoux, and P. Frey. Block-structured hexahedral meshes for cad models using 3d frame fields. Procedia Engineering, 82:59 – 71, 2014. 23rd International Meshing Roundtable (IMR23)

    2014

  12. [19]

    Kowalski, F

    N. Kowalski, F. Ledoux, and P. Frey. Smoothness driven frame field generation for hexahedral meshing. Computer-Aided Design, 72:65 – 77, 2016. 23rd International Meshing Roundtable Special Issue: Advances in Mesh Generation

    2016

  13. [20]

    Y. Li, Y. Liu, W. Xu, W. Wang, and B. Guo. All-hex meshing using singularity-restricted field. ACM Trans. Graph., 31(6):177:1–177:11, Nov. 2012

    2012

  14. [21]

    H. Liu, P. Zhang, E. Chien, J. Solomon, and D. Bommes. Singularity-constrained octahedral fields for hexa- hedral meshing. ACM Trans. Graph., 37(4):93:1–93:17, July 2018

    2018

  15. [22]

    N. J. Mottram and C. J. Newton. Introduction to Q-tensor theory. arXiv preprint arXiv:1409.3542 , 2014

    2014 arXiv

  16. [23]

    Nieser, U

    M. Nieser, U. Reitebuch, and K. Polthier. Cubecover parameterization of 3d volumes. Computer Graphics Forum, 30(5):1397–1406, 2011

    2011

  17. [24]

    Palacios and E

    J. Palacios and E. Zhang. Rotational symmetry field design on surfaces. ACM Trans. Graph., 26(3), July 2007

    2007

  18. [26]

    Ray and D

    N. Ray and D. Sokolov. Robust polylines tracing for n-symmetry direction field on triangulated surfaces. ACM Trans. Graph., 33(3):30:1–30:11, June 2014

    2014

  19. [27]

    N. Ray, D. Sokolov, and B. L´ evy. Practical 3d frame field generation.ACM Transactions on Graphics (TOG) , 35(6):233, 2016

    2016

  20. [28]

    N. Ray, B. Vallet, W. C. Li, and B. L´ evy. N-symmetry direction field design.ACM Trans. Graph., 27(2):10:1– 10:13, May 2008

    2008

  21. [29]

    E. Robeva. Orthogonal decomposition of symmetric tensors. SIAM Journal on Matrix Analysis and Applica- tions, 37(1):86–102, 2016

    2016

  22. [30]

    Solomon, A

    J. Solomon, A. Vaxman, and D. Bommes. Boundary element octahedral fields in volumes. ACM Trans. Graph., 36(3), May 2017

    2017

  23. [31]

    Vaxman, M

    A. Vaxman, M. Campen, O. Diamanti, D. Panozzo, D. Bommes, K. Hildebrandt, and M. Ben-Chen. Directional Field Synthesis, Design, and Processing. COMPUTER GRAPHICS FORUM , 35(2):545–572, MAY 2016

    2016

  24. [32]

    Viertel and B

    R. Viertel and B. Osting. An approach to quad meshing based on harmonic cross-valued maps and the ginzburg– landau theory. SIAM Journal on Scientific Computing , 41(1):A452–A479, 2019

    2019

  25. [33]

    Viertel, M

    R. Viertel, M. L. Staten, and F. Ledoux. Analysis of non-meshable automatically generated frame fields. Technical report, Sandia National Lab.(SNL-NM), Albuquerque, NM (United States), 2016. Department of Mathematics, The University of Akron, Akron, OH 44325, USA E-mail address...

    2016

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.