pith. sign in

arxiv: 1907.07550 · v1 · pith:WZK73FLSnew · submitted 2019-07-17 · 🧮 math.NA · cs.NA

Geometric subdivision and multiscale transforms

Johannes Wallner This is my paper

Pith reviewed 2026-05-24 20:13 UTC · model grok-4.3

classification 🧮 math.NA cs.NA
keywords geometric subdivisionmultiscale transformsRiemannian manifoldspositive definite matricessymmetry preservationrefinement proceduresconvergencesmoothness
0
0 comments X

The pith

Any procedure on data must respect the symmetries and geometric structure of that data.

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

The paper establishes an axiom that every operation applied to data, including averages, refinement, and multiscale transforms, must preserve the intrinsic nature and symmetries of the data. It examines how this requirement can be met by working directly with geometric structures such as metric spaces, Riemannian manifolds, and groups rather than forcing data into Euclidean coordinates. A concrete case is the natural Riemannian metric on the space of positive definite matrices, which yields intrinsic averages and operations. The discussion then turns to subdivision schemes and multiresolution transforms defined on these structures, summarizing existing results on their convergence and smoothness. A reader would care because procedures that ignore geometry can distort properties that the data are meant to carry.

Core claim

The author claims that refinement procedures and multiresolution transforms must be constructed from intrinsic operations on geometric structures (metric spaces, Riemannian manifolds, groups) so that they respect data symmetries, illustrates the principle with the Riemannian metric on positive definite matrices, and reviews the current state of knowledge concerning convergence and smoothness of the resulting subdivision schemes.

What carries the argument

Intrinsic operations derived from geometric structures such as the Riemannian metric on positive definite matrices, used to define averages and refinement rules that preserve symmetries.

If this is right

  • Averages and subdivisions defined via the Riemannian metric on positive definite matrices preserve positive definiteness and other matrix properties.
  • Subdivision schemes on manifolds converge when the refinement rules are constructed from the intrinsic geometry.
  • Multiscale transforms built this way maintain the symmetries of the original data space.
  • Smoothness of the limit surface or curve depends on the choice of geometrically consistent averaging.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same symmetry requirement would apply to data-driven methods such as interpolation on Lie groups used in robotics or animation.
  • One could test whether Euclidean approximations systematically introduce artifacts on spheres or hyperbolic spaces that intrinsic methods avoid.
  • The approach suggests examining whether current multiresolution tools in scientific computing already satisfy or violate the axiom on specific manifolds.

Load-bearing premise

That metric spaces, Riemannian manifolds, and groups supply the right setting in which to define operations that respect the nature of the data.

What would settle it

A concrete subdivision scheme or average defined without reference to the manifold structure that nonetheless produces limits with the expected convergence rate and smoothness on a Riemannian manifold or group.

Figures

Figures reproduced from arXiv: 1907.07550 by Johannes Wallner.

Figure 1
Figure 1. Figure 1: Subdivision by projection in the motion group R 3 o O3. A 4-periodic sequence pi = (ci , ui) of positions of a rigid body is defined by the center of mass ci , and an orientation ui ∈ O3. Both components undergo subdivision w.r.t. the interpolatory four-point rule S, where the matrix part is subsequently projected back onto O3 in an invariant manner. • Subdivision using projections. If M is a surface embed… view at source ↗
Figure 2
Figure 2. Figure 2: Geodesic corner-cutting rules are among those where convergence is not difficult to show. These images show Chaikin’s rule S (2), with the original data in red, and the result of subdivision as a yellow geodesic polygon. Sp2i = pi and Sp2i+1 = −ωpi−1 + ( 1 2 + ω)pi + ( 1 2 + ω)pi+1 − ωpi+2 = pi + pi+1 2 − ω  pi−1 − pi + pi+1 2  − ω  pi+2 − pi + pi+1 2  . In the special case ω = 1 16 , the point Sp2i+1 … view at source ↗
Figure 3
Figure 3. Figure 3: Subdivision rules Spj = avgF (aj−2i ; pi) based on the Fr´echet mean operating on sequences on the unit sphere. The images visualize the interpolatory 4-point rule (left) and a rule without any special properties. We show the coefficient sequence aj and the bound on δ(p) which ensures convergence. This result is satisfying because it allows us to infer convergence from a condition which is well known in th… view at source ↗
Figure 4
Figure 4. Figure 4: Left: Hermite data (pi , vi) in R 2 and the result of one round subdivision by a linear Hermite rule S. Center: Limit curve f (f 0 is not shown). Right: Hermite data (pi , vi) in the group SO3, and the limit curve generated by a group version of S. Points pi ∈ SO3 and tangent vectors vi ∈ TpiSO3 are visualized by means of their action on a spherical triangle. These figures appeared in [42] (reprinted with … view at source ↗
Figure 5
Figure 5. Figure 5: Here data pi in the unit sphere Σ2 and Pos3-valued data qi are visualized by plac￾ing the ellipsoid with equation x T qix = 1 in the point pi ∈ Σ 2 . Both data undergo iterative refinement by means of a Riemannian version S of the Doo-Sabin subdivision rule. For given initial data p, q which have the combinatorics of a cube, the four images show S jp and S j q, for q = 1, 2, 3, 4 (from left). The correspon… view at source ↗
Figure 6
Figure 6. Figure 6: The decomposition and re￾construction chains of operations in a geometric multiscale decomposition based on upscaling and downscaling S, D for points, and upscaling and downscaling R, Q for detail vectors [PITH_FULL_IMAGE:figures/full_fig_p020_6.png] view at source ↗
read the original abstract

Any procedure applied to data, and any quantity derived from data, is required to respect the nature and symmetries of the data. This axiom applies to refinement procedures and multiresolution transforms as well as to more basic operations like averages. This chapter discusses different kinds of geometric structures like metric spaces, Riemannian manifolds, and groups, and in what way we can make elementary operations geometrically meaningful. A nice example of this is the Riemannian metric naturally associated with the space of positive definite matrices and the intrinsic operations on positive definite matrices derived from it. We disucss averages first and then proceed to refinement operations (subdivision) and multiscale transforms. In particular, we report on the current knowledge as regards convergence and smoothness.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

0 major / 1 minor

Summary. The manuscript is an expository survey chapter that states an axiom requiring any data-processing procedure (including averages, refinement schemes, and multiresolution transforms) to respect the symmetries and intrinsic geometry of the underlying data. It reviews standard geometric settings—metric spaces, Riemannian manifolds, and groups—and shows how to equip them with intrinsic operations, using the Riemannian metric on the manifold of positive definite matrices as a running example. The chapter then surveys known constructions of geometric subdivision schemes and multiscale transforms, reporting existing results on their convergence and smoothness without presenting new theorems.

Significance. As a survey, the chapter offers a coherent framing of geometric subdivision and multiresolution methods under a single symmetry-respecting principle. Its value lies in collecting and organizing known convergence and smoothness results from the literature into one narrative, which may serve as a useful reference for researchers in numerical analysis and geometric modeling. No new derivations, proofs, or empirical validations are provided, so the significance is that of synthesis rather than original contribution.

minor comments (1)
  1. [Abstract] Abstract: 'disucss' is a typographical error for 'discuss'.

Simulated Author's Rebuttal

0 responses · 0 unresolved

We thank the referee for the positive assessment of the manuscript as an expository survey chapter and for the recommendation of minor revision. The report accurately characterizes the work as a synthesis of existing results on symmetry-respecting geometric operations, subdivision schemes, and multiscale transforms, without new theorems.

Circularity Check

0 steps flagged

No significant circularity; expository survey with external citations

full rationale

The paper is a survey chapter presenting an axiom about respecting data symmetries and then surveying known constructions on standard geometric structures (metric spaces, Riemannian manifolds, groups). It reports existing results on convergence and smoothness without offering new derivations, predictions, or theorems. No load-bearing steps reduce to self-citation chains or fitted inputs by construction; all technical content is attributed to external literature. This matches the default expectation of a self-contained expository text.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

This is a survey chapter; the ledger records background assumptions from differential geometry and analysis that the discussion relies upon rather than new postulates introduced by the author.

axioms (2)
  • domain assumption Data possesses symmetries that any derived quantity must respect.
    Opening sentence of the abstract; treated as an axiom for the entire discussion.
  • domain assumption Riemannian metrics on manifolds such as the space of positive definite matrices induce intrinsic, geometrically meaningful operations.
    Presented as the motivating example for making averages and refinement intrinsic.

pith-pipeline@v0.9.0 · 5632 in / 1220 out tokens · 18139 ms · 2026-05-24T20:13:58.821988+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

62 extracted references · 62 canonical work pages · 1 internal anchor

  1. [1]

    Ballmann, Lectures on spaces of nonpositive curvature , Birkh¨ auser, 1995

    W. Ballmann, Lectures on spaces of nonpositive curvature , Birkh¨ auser, 1995

    work page 1995

  2. [2]

    225, Springer-Verlag, 2004

    Daniel Bump, Lie groups, Graduate Texts in Mathematics, vol. 225, Springer-Verlag, 2004

    work page 2004

  3. [3]

    Catmull and J

    E. Catmull and J. Clark, Recursively generated b-spline surfaces on arbitrary topological meshes , Computer-Aided Design 10 (1978), 350–355

    work page 1978

  4. [4]

    Cavaretta, Wolfgang Dahmen, and Charles A

    Albert S. Cavaretta, Wolfgang Dahmen, and Charles A. Michelli, Stationary subdivision, Memoirs AMS, vol. 93, American Math. Soc., 1991

    work page 1991

  5. [5]

    Chaikin, An algorithm for high speed curve generation , Comp

    G. Chaikin, An algorithm for high speed curve generation , Comp. Graphics Image Processing 3 (1974), 346–349. GEOMETRIC SUBDIVISION AND MULTISCALE TRANSFORMS 24

    work page 1974

  6. [6]

    Sabin, Behaviour of recursive division surfaces near extraordinary points, Computer- Aided Design 10 (1978), 356–360

    D.Doo and M. Sabin, Behaviour of recursive division surfaces near extraordinary points, Computer- Aided Design 10 (1978), 356–360

    work page 1978

  7. [7]

    de Rham, Sur quelques fonctions diff´ erentiables dont toutes les valeurs sont des valeurs critiques, Celebrazioni Archimedee del Sec

    G. de Rham, Sur quelques fonctions diff´ erentiables dont toutes les valeurs sont des valeurs critiques, Celebrazioni Archimedee del Sec. XX (Siracusa, 1961), Vol. II, Edizioni “Oderisi”, Gubbio, 1962, pp. 61–65

    work page 1961

  8. [8]

    Deslauriers and S

    G. Deslauriers and S. Dubuc, Symmetric iterative interpolation processes, Constr. Approx. 5 (1986), 49–68

    work page 1986

  9. [9]

    do Carmo, Riemannian geometry, Birkh¨ auser Verlag, 1992

    Manfredo P. do Carmo, Riemannian geometry, Birkh¨ auser Verlag, 1992

    work page 1992

  10. [10]

    D. L. Donoho, Interpolating wavelet transforms, Tech. report,http://www-stat.stanford.edu/donoho/ Reports/1992/interpol.pdf, 1992

    work page 1992

  11. [11]

    Approximation and Computation

    , Wavelet-type representation of Lie-valued data, 2001, Talk at the IMI “Approximation and Computation” meeting, May 12–17, 2001, Charleston, South Carolina

    work page 2001

  12. [12]

    Duchamp, G

    T. Duchamp, G. Xie, and T. Yu, On a new proximition condition for manifold-valued subdivision schemes, Approximation Theory XIV: San Antonio 2013 (Gregory E. Fasshauer and Larry L. Schu- maker, eds.), Springer Proceedings in Mathematics & Statistics, vol. 83, Springer, 2014, pp. 65–79

    work page 2013

  13. [13]

    , A necessary and sufficient proximity condition for smoothness equivalence of nonlinear subdivision schemes, Found. Comp. Math. (2015), 1–46

    work page 2015

  14. [14]

    Algorithms 77 (2018), 361–379

    Tom Duchamp, Gang Xie, and Thomas Yu, Smoothing nonlinear subdivision schemes by averaging , Numer. Algorithms 77 (2018), 361–379. MR 3748375

    work page 2018

  15. [15]

    R. Dyer, G. Vegter, , and M. Wintraecken, Barycentric coordinate neighbourhoods in Riemannian manifolds, arxiv https://arxiv.org/abs/1606.01585, 2016, 15pp

    work page internal anchor Pith review Pith/arXiv arXiv 2016

  16. [16]

    , Barycentric coordinate neighbourhoods in Riemannian manifolds , Extended Abstracts, SoCG Young Researcher Forum, 2016, pp. 1–2

    work page 2016

  17. [17]

    Dyn, Subdivision schemes in computer-aided geometric design , Advances in Numerical Analysis Vol

    N. Dyn, Subdivision schemes in computer-aided geometric design , Advances in Numerical Analysis Vol. II (W. A. Light, ed.), Oxford Univ. Press, 1992, pp. 36–104

    work page 1992

  18. [18]

    N. Dyn, J. Gregory, and D. Levin, A four-point interpolatory subdivision scheme for curve design , Comput. Aided Geom. Des. 4 (1987), 257–268

    work page 1987

  19. [19]

    Dyn and D

    N. Dyn and D. Levin, Subdivision schemes in geometric modelling , Acta Numer. 11 (2002), 73–144

    work page 2002

  20. [20]

    of Compu- tation 86 (2017), 375–395

    Nira Dyn and Nir Sharon, A global approach to the refinement of manifold data , Math. of Compu- tation 86 (2017), 375–395

    work page 2017

  21. [21]

    , Manifold-valued subdivision schemes based on geodesic inductive averaging , J. Comput. Appl. Math. 311 (2017), 54–67

    work page 2017

  22. [22]

    American Math

    Oliver Ebner, Convergence of iterative schemes in metric spaces , Proc. American Math. Soc. 141 (2013), 677–686

    work page 2013

  23. [23]

    , Stochastic aspects of refinement schemes on metric spaces , SIAM J. Num. Anal. 52 (2014), 717–734

    work page 2014

  24. [24]

    P. T. Fletcher and S. Joshi, Principal geodesic analysis on symmetric spaces: Statistics of diffusion tensors, Computer Vision and Mathematical Methods in Medical and Biomedical Image Analysis (M. Sonka et al., eds.), LNCS, no. 3117, Springer, 2004, pp. 87–98

    work page 2004

  25. [25]

    Philipp Grohs, Smoothness analysis of subdivision schemes on regular grids by proximity , SIAM J. Num. Analysis 46 (2008), 2169–2182

    work page 2008

  26. [26]

    , Smoothness equivalence properties of univariate subdivision schemes and their projection analogues, Num. Math. 113 (2009), 163–180

    work page 2009

  27. [27]

    , Smoothness of interpolatory multivariate subdivision in Lie groups , IMA J. Numer. Anal. 27 (2009), 760–772

    work page 2009

  28. [28]

    , Approximation order from stability of nonlinear subdivision schemes , J. Approx. Th. 162 (2010), 1085–1094

    work page 2010

  29. [29]

    , A general proximity analysis of nonlinear subdivision schemes , SIAM J. Math. Anal. 42 (2010), no. 2, 729–750

    work page 2010

  30. [30]

    , Stability of manifold-valued subdivision and multiscale transforms , Constr. Approx. 32 (2010), 569–596. GEOMETRIC SUBDIVISION AND MULTISCALE TRANSFORMS 25

    work page 2010

  31. [31]

    Neamtu and L

    Philipp Grohs and Johannes Wallner, Log-exponential analogues of univariate subdivision schemes in Lie groups and their smoothness properties , Approximation Theory XII: San Antonio 2007 (M. Neamtu and L. L. Schumaker, eds.), Nashboro Press, 2008, pp. 181–190

    work page 2007

  32. [32]

    , Interpolatory wavelets for manifold-valued data , Appl. Comput. Harmon. Anal. 27 (2009), 325–333

    work page 2009

  33. [33]

    Franklin Inst

    , Definability and stability of multiscale decompositions for manifold-valued data , J. Franklin Inst. 349 (2012), 1648–1664

    work page 2012

  34. [34]

    thesis, FU Berlin, 2015

    Hanne Hardering, Intrinsic discretization error bounds for geodesic finite elements , Ph.D. thesis, FU Berlin, 2015

    work page 2015

  35. [35]

    Sigurdur Helgason, Differential geometry, Lie groups, and symmetric spaces , Academic Press, 1978

    work page 1978

  36. [36]

    Svenja H¨ uning and Johannes Wallner,Convergence analysis of subdivision processes on the sphere , submitted, 2019

    work page 2019

  37. [37]

    , Convergence of subdivision schemes on Riemannian manifolds with nonpositive sectional curvature, Adv. Comput. Math. 45 (2019), 1689–1709

    work page 2019

  38. [38]

    Uri Itai and Nir Sharon, Subdivision schemes for positive definite matrices , Found. Comput. Math. 13 (2013), no. 3, 347–369

    work page 2013

  39. [39]

    Pure Appl

    Hermann Karcher, Riemannian center of mass and mollifier smoothing , Comm. Pure Appl. Math. 30 (1977), 509–541

    work page 1977

  40. [40]

    Matrix Anal

    Maher Moakher, A differential geometric approach to the geometric mean of symmetric positive- definite matrices, SIAM J. Matrix Anal. Appl. 26 (2005), no. 3, 735–747

    work page 2005

  41. [41]

    Caroline Moosm¨ uller,C 1 analysis of Hermite subdivision schemes on manifolds , SIAM J. Num. Analysis 54 (2016), 3003–3031

    work page 2016

  42. [42]

    Caroline Moosm¨ uller,Hermite subdivision on manifolds via parallel transport , Adv. Comput. Math. 43 (2017), 1059–1074

    work page 2017

  43. [43]

    J¨ org Peters and Ulrich Reif,Subdivision surfaces, Springer, 2008

    work page 2008

  44. [44]

    Aided Geom

    Ulrich Reif, A unified approach to subdivision algorithms near extraordinary vertices, Comput. Aided Geom. Des. 12 (1995), no. 2, 153–174

    work page 1995

  45. [45]

    Riesenfeld, On Chaikin’s algorithm, IEEE Comp

    R. Riesenfeld, On Chaikin’s algorithm, IEEE Comp. Graph. Appl. 4 (1975), no. 3, 304–310

    work page 1975

  46. [46]

    Sturm, Nonlinear martingale theory for processes with values in metric spaces of nonpositive curvature, Ann

    K.-T. Sturm, Nonlinear martingale theory for processes with values in metric spaces of nonpositive curvature, Ann. Probab. 30 (2002), 1195–1222

    work page 2002

  47. [47]

    , Probability measures on metric spaces of nonpositive curvature , Heat kernels and analysis on manifolds, graphs, and metric spaces, AMS, 2003, pp. 357–390

    work page 2003

  48. [48]

    Stodden, David L

    Inam Ur Rahman, Iddo Drori, Victoria C. Stodden, David L. Donoho, and Peter Schr¨ oder,Multiscale representations for manifold-valued data , Multiscale Mod. Sim. 4 (2005), 1201–1232

    work page 2005

  49. [49]

    Johannes Wallner, Smoothness analysis of subdivision schemes by proximity , Constr. Approx. 24 (2006), no. 3, 289–318

    work page 2006

  50. [50]

    , On convergent interpolatory subdivision schemes in Riemannian geometry, Constr. Approx. 40 (2014), 473–486

    work page 2014

  51. [51]

    Aided Geom

    Johannes Wallner and Nira Dyn, Convergence andC 1 analysis of subdivision schemes on manifolds by proximity, Comput. Aided Geom. Design 22 (2005), 593–622

    work page 2005

  52. [52]

    Johannes Wallner, Esfandiar Nava Yazdani, and Philipp Grohs, Smoothness properties of Lie group subdivision schemes, Multiscale Modeling and Simulation 6 (2007), 493–505

    work page 2007

  53. [53]

    Johannes Wallner, Esfandiar Nava Yazdani, and Andreas Weinmann, Convergence and smoothness analysis of subdivision rules in Riemannian and symmetric spaces , Adv. Comp. Math 34 (2011), 201–218

    work page 2011

  54. [54]

    Graphics 25 (2006), no

    Johannes Wallner and Helmut Pottmann, Intrinsic subdivision with smooth limits for graphics and animation, ACM Trans. Graphics 25 (2006), no. 2, 356–374

    work page 2006

  55. [55]

    Andreas Weinmann, Nonlinear subdivision schemes on irregular meshes, Constr. Approx. 31 (2010), 395–415

    work page 2010

  56. [56]

    , Interpolatory multiscale representation for functions between manifolds , SIAM J. Math. Analysis 44 (2012), 172–191

    work page 2012

  57. [57]

    , Subdivision schemes with general dilation in the geometric and nonlinear setting, J. Approx. Th. 164 (2012), 105–137. GEOMETRIC SUBDIVISION AND MULTISCALE TRANSFORMS 26

    work page 2012

  58. [58]

    M. Welk, J. Weickert, F. Becker, C. Schn¨ orr, and B. Burgeth C. Feddern and, Median and related local filters for tensor-valued images , Signal Processing 87 (2007), 291–308

    work page 2007

  59. [59]

    Xie and T

    G. Xie and T. P.-Y. Yu, Smoothness equivalence properties of manifold-valued data subdivision schemes based on the projection approach , SIAM J. Numer. Anal. 45 (2007), 1200–1225

    work page 2007

  60. [60]

    , Smoothness equivalence properties of general manifold-valued data subdivision schemes , Multiscale Mod. Sim. 7 (2008), 1073–1100

    work page 2008

  61. [61]

    , Approximation order equivalence properties of manifold-valued data subdivision schemes , IMA J. Num. Analysis 32 (2011), 687–700

    work page 2011

  62. [62]

    Yu, Smoothness equivalence properties of interpolatory Lie group sub- division schemes, IMA J

    Gang Xie and Thomas P.-Y. Yu, Smoothness equivalence properties of interpolatory Lie group sub- division schemes, IMA J. Numer. Anal. 30 (2010), 731–750. MR 2670112 Johannes W allner. TU Graz, Kopernikusgasse 24, 80180 Graz, Austria, email j.wallner@tugraz.at

    work page 2010