Pith. sign in

REVIEW 3 major objections 6 minor 26 references

Detection of symmetry using a crystallographic image processing algorithm

T0 review · 3 major / 6 minor · reviewed 2026-08-14 · deepseek-v4-flash

Pith's one-line read The paper claims that lower phase residuals in a crystallographic image processing search identify the correct plane symmetry group of a periodic 2D image, and that phase origin maps reveal hidden and broken symmetry elements.

desk verdict A solid teaching-oriented demonstration of known CIP methods on Escher woodcuts, with an overstated headline claim about automated symmetry detection that the paper itself partially contradicts. read the letter →

arxiv 1908.07932 v1 pith:2JB4LGFF submitted 2019-08-08 physics.pop-ph physics.ed-ph

classification physics.pop-phphysics.ed-ph
keywords symmetrydetectionplanegroupdeterminationphaseresidualoriginmapcrystallographicimageprocessingblack-and-whitecolorFourieranalysis
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 argues that a crystallographic image processing workflow—Fourier transform, indexing, symmetrization, and residual scoring—can automatically identify the plane symmetry group of any periodic two-dimensional image. Its central claim is that the plane group with the lowest phase residual is the most likely true symmetry, and that the phase origin map exposes symmetry elements hidden by coloring or broken by small defects. Demonstrations on four images, including a famous artist's tessellations and an 'impossible figure', show the method ranking candidate plane groups and locating mirror, glide, and rotation features that are not obvious in real space. If true, the approach turns symmetry detection into a quantitative Fourier-based task rather than an act of visual inspection.

What carries the argument

The central machinery is the phase residual functional and its visualization as a phase origin map (POM). For a candidate plane group, each observed Fourier component is compared with a 'symmetrized phase' obtained by averaging all symmetry-related components; the amplitude-weighted sum of phase differences yields a residual. Shifting the assumed origin by $(\varphi_x,\varphi_y)$ phases each component by $\varphi_x h + \varphi_y k$, and plotting the residual over all shifts produces the POM. In the paper's interpretation, deep minima mark rotation axes and dark lines mark mirror or glide lines, so the map serves both as an origin refiner and as a symmetry detector.

What would settle it

Take a synthetic image with an exactly known plane group, origin, and a single known mirror line; run the same pipeline; if another plane group scores a lower phase residual, or the phase origin map's dark line is displaced from the known mirror position (or absent), the central claim is refuted. Repeat over all 17 plane groups with controlled noise.

Watch

Extended reading notes

Core claim

On its own terms, the paper discovers that the phase statistics of a periodic image's Fourier components carry enough symmetry information to classify its plane group and to localize symmetry elements. The symmetrized phase formula averages symmetry-related Fourier components, and the phase residual measures how far observed phases deviate from the relations imposed by each of the 17 plane groups; the lower the residual, the more likely the group. The phase origin map, built by plotting the residual as a function of origin shifts, shows dark lines and minima at positions corresponding to mirror/glide lines and rotation axes. By inspecting these maps, the author claims to detect symmetry elements that are not apparent in real space: a broken mirror in an 'impossible' tessellation, black-and-white glide symmetries, and color-hidden 6-fold axes.

Load-bearing premise

The method assumes that lower phase residuals directly indicate the correct plane group and that dark lines and minima in the phase origin map are faithful tracings of mirror, glide, and rotation elements in real space; the paper supports this only by visual comparison on four selected images rather than an independent ground-truth test.

Editorial extensions

If this is right

  • Automated screening can rank the 17 plane groups for a periodic 2D image, with the lowest phase residual indicating the most probable group.
  • Phase origin maps can reveal 'broken' symmetry elements—mirror lines that almost hold—and 'hidden' symmetry elements such as color-changing glide lines.
  • Applying an edge-detection filter to a colored periodic pattern strips color and lets the same algorithm detect the color-blind supergroup.
  • Systematically absent reflections, quantified by the Fo/Fe ratio, help distinguish plane groups that share phase relations and signal imperfect group membership.

Reading between the lines

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

  • A natural next test would be a ground-truth benchmark: generate synthetic periodic images in each of the 17 plane groups with known origins and controlled noise, then measure how often the lowest phase residual identifies the correct group and how accurately POM minima locate symmetry elements.
  • The same scoring could be applied to other 'structured' images—quasicrystal tilings, moiré patterns, or biological tissue sections—where local periodic order is suspected but a human eye cannot decide the symmetry group.
  • Because the residual is amplitude-weighted, the method's sensitivity to broken symmetry is likely dominated by strong low-order Fourier components; small but symmetry-breaking details with weak high-frequency content may be systematically under-weighted.
  • Comparing residuals between a group and its subgroups, as the paper does, could be formalized into a statistical model-selection criterion rather than a heuristic ranking.
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

3 major / 6 minor

Summary. The paper applies crystallographic image processing (CIP), originally developed for HRTEM, to digitized 2D images such as Escher woodcuts. It computes phase and amplitude residuals for candidate plane groups, uses the phase origin map (POM) to locate symmetry elements, and enforces or refines symmetry to reconstruct images. Four examples are used to argue that residuals rank candidate plane groups, that the POM reveals hidden or broken symmetry elements, and that Sobel-filtered colorblind versions of images expose supergroups. The central claims are that lower phase residuals indicate higher likelihood of the correct plane group (Section 2.2) and that POM features such as dark lines and minima correspond to mirror/glide lines and rotation axes (Section 3.4 and Section 4).

Significance. If fully supported, the method would provide an automated, quantitative way to assign plane groups to periodic 2D patterns and to identify symmetry elements that are not obvious in real space, with didactic value for crystallography teaching. The mathematical basis in Zou and Hovmöller's residual formulas is standard, and the four worked examples are internally consistent with visual inspection of the figures. However, the advertised residual-based classifier is contradicted by the paper's own tables, and the POM interpretation is not validated against a ground truth or synthetic artifact control. The absence of code, data, and error bars prevents independent verification of the quantitative residuals. These issues make the central claims currently unsupported as stated, though the underlying idea may be salvageable with a revised decision rule and validation.

major comments (3)
  1. [Section 2.2, Eq. (2) and Table 1, Section 4.4] The claim that 'the lower the residuals are for a given plane symmetry group, the higher is the likelihood that this group is the right one' is directly contradicted by the paper's own data. In Table 1, the image is assigned to p4gm, but p4gm has phase residual 46.5, while the subgroups p2 (24.2) and p1g1/p11g (24.1) have much lower residuals. In Section 4.4, the correct group p2 has residual 20.56, but p3 has 18.85. The author's recourse to 'look for the highest symmetry possible' and to compare refined images (Section 4.1) is an additional criterion that is not derived from Eq. (2) and is not formalized. As written, the residual ranking therefore does not provide an automated symmetry classifier; it must be replaced or supplemented by a clear rule that accounts for subgroup/supergroup relations, for example by comparing residuals only among groups that are not related by subgroup inclusion or by adding a penalty for additional symmetry relations.
  2. [Section 3.4 and Sections 4.2-4.4] The interpretation of phase origin map features as 'hidden' or 'broken' symmetry elements is not validated by any independent ground-truth test. The paper asserts, for instance, that dark lines in the POM correspond to mirror lines, dimmer lines to glide lines, and dark dots to rotation axes, but the only support is visual alignment on four selected images. No synthetic images with known symmetry and controlled noise, no comparison with a full 17-plane-group benchmark, and no test of whether windowing, finite image size, or peak-search truncation can generate similar POM features are provided. Without such a control, the central inference that POM features directly encode real-space symmetry is an unsupported premise, and the conclusions about broken symmetry in Example 4.2 and hidden color symmetry in Examples 4.3 and 4.4 do not follow.
  3. [Section 3.1 and Tables 1-4] The paper provides no code, data, or detailed numerical output beyond a few residual values, so the residual tables cannot be independently reproduced or checked. Given that the central claim is quantitative and that the residual differences between groups can be small (for example, 20.56 vs. 18.85 in Section 4.4), error bars or a reproducibility statement are necessary. The reader cannot tell whether the reported residuals are stable under changes in windowing, peak search, and basis selection, which are all described as involving user choices (Section 3.1 and 3.2).
minor comments (6)
  1. [Abstract and Section 1] The abstract says 'an automated method to quantify and detect symmetry elements,' but the final decision in each example relies on visual comparison of refined images with the original; the criterion should be clarified in the abstract.
  2. [Section 2.2] The definition of the R-factor in ALLSPACE appears as an inline formula without equation numbering; renumbering or referencing it would improve readability.
  3. [Section 3.1] The sentence about the auto cross-correlation of the power spectrum reveals 'previously invisible' peaks, but the mechanism is not explained; a brief mathematical description would help.
  4. [Section 3.2, Eq. (5)] The notation in Eq. (5) is hard to parse because the rounding operator and the summation limits are not defined consistently; please define the index set over which the sum runs.
  5. [Section 4.3] The phrase 'pgy (p11g)' uses a nonstandard symbol 'pgy'; the full standard symbol p1g1 is introduced later, so the notation should be made consistent.
  6. [Section 4.4] The statement 'p3 and p6 are very close to p2' is ambiguous; please give the exact residual differences and a statistical or heuristic threshold for 'close'.

Circularity Check

1 steps flagged · score 6.0 of 10

Partial circularity: the residual-based likelihood ranking is a self-consistency fit that favors less constrained groups; the paper's own tables contradict it, and the final group selection depends on an extra 'highest symmetry' rule.

  1. self definitional [Section 2.2, Eqs. (1)-(2); Section 4.1, Table 1]
    "If a FC is not related to any other FC by the symmetry (except by the Friedel’s law), then: φsym(hk) = φobs(hk). ... The lower the residuals φRes are for a given plane symmetry group, the higher is the likelihood that this group is the right one for the pattern under investigation."

    Eq. (1) defines the symmetrized phase as a weighted average of the observed phases, so for any group with no symmetry relations φsym equals φobs and the phase residual (2) is identically zero by construction. The residual therefore measures self-consistency of the data with their own average, not agreement with an externally derived group model; fewer relations always make the fit trivially easier. The paper's own Table 1 confirms this: the correct p4gm has residual 46.5 while its subgroups p2 and p1g1/p11g have about 24.1-24.2, and Section 4.1 then adds an extra 'we look for the highest symmetry possible' criterion, which is not contained in Eq. (2).

full rationale

The main derivation chain—Fourier transform, peak search, symmetrized-phase residual, and phase origin map—is a direct numerical transform of the image data, and the POM interpretation is not definitionally circular. However, the methodological core in Section 2.2 is partially circular: φsym is a weighted average of the observed phases, and the paper explicitly sets φsym=φobs for unconstrained reflections, so Eq. (2) is minimized by construction for groups with fewer symmetry relations (identically zero for p1). Calling the resulting residual a 'likelihood' that the tested group is correct is therefore a fitted-input-as-prediction: the residual is a self-consistency score, not an independent test. The paper's own Table 1 and Section 4.4 admit that subgroups/supergroups have lower/higher residuals, and Section 4.1 introduces the additional 'highest symmetry possible' and refined-image comparison to escape the tautology. That extra criterion is not derived from Eqs. (1)-(2), so the central automated-classification claim is partially reduced by construction. The self-citations [20]-[22] are to the author's software implementation, not to a load-bearing theorem; the underlying CIP algorithm is externally attributed to [9] and [19]. The POM hidden/broken-symmetry readings are interpretive and are supported only by visual alignment on four selected images, which is a validation weakness rather than a circular step. Overall score 6 because one central predictive claim reduces by construction, while the broader method retains independent content.

Assumptions & free parameters 2 free parameters · 6 assumptions · 0 invented entities

The central claim rests on standard Fourier and group-theoretic relations, on prior crystallographic image processing formulas (Zou-Hovmoller, CRISP), and on a more fragile interpretive rule: that phase origin map features map to real-space symmetry elements. No new free entities are introduced. The algorithmic parameters, such as basis weights and peak cutoff, are not fully specified and act as hidden free parameters.

free parameters (2)
  • Basis-selection weighting factors = not quantified in the paper
    Section 3.2 states that weights make rectangular and hexagonal lattices preferable over oblique ones; the exact weights are never given, yet they influence which basis is chosen and therefore all subsequent residual values.
  • Peak-search and Fourier-component cutoff = not specified; examples use 8264, 1680, 562 components
    The paper reports the number of collected Fourier components for each example but does not state the threshold that stops peak collection. Residuals and Fo/Fe ratios depend on this cutoff.
assumptions (6)
  • standard math The Fourier transform of a periodic image is a discrete set of Fourier components, and plane-group symmetry imposes exact relationships between amplitudes and phases of symmetry-related components.
    Stated in Sections 2.1 and 3.3; underlies every residual computation. This is standard Fourier and group theory for periodic functions.
  • domain assumption The symmetrized phase formula (Eq. 1) and phase residual functional (Eq. 2) from Zou and Hovmoller correctly measure how well an image conforms to a candidate plane group.
    Adopted from prior crystallographic image processing literature (Ref. [9]); the paper uses these formulas as the core detection metric without independent validation in this new image domain.
  • standard math Plane-group phase relations tabulated in International Tables for Crystallography Vol. B are complete and correct for all wallpaper groups considered.
    Used in Section 3.3 to enumerate symmetry-related Fourier components; this is the established reference in the field.
  • ad hoc to paper Lower phase residuals correspond to higher likelihood of the correct plane group, and, among low-residual groups, the one with highest symmetry that reproduces the image upon refinement is the true group.
    Invoked in Section 4.1 to select p4gm over its subgroups even though p4gm has higher residuals; no statistical test is given.
  • domain assumption The phase origin map, computed as phase residual over shifts of the origin, has minima and extended features that correspond to real-space positions of symmetry elements.
    Central to the hidden and broken symmetry claims in Examples 4.2 to 4.4; the correspondence is inferred by visual inspection, not proven.
  • domain assumption Applying a Sobel edge filter yields a colorblind image whose plane group is the grey supergroup of the colored pattern.
    Used in Examples 4.3 and 4.4 to reveal color symmetry; assumes edge filtering removes color information without destroying the geometry.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Detection of symmetry using a crystallographic image processing algorithm." pith.science (2026). https://pith.science/paper/2JB4LGFF

@misc{pith2026190807932,
  author       = {Pith},
  title        = {Pith review of: Detection of symmetry using a crystallographic image processing algorithm},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/2JB4LGFF}},
  note         = {Machine review of arXiv:1908.07932}
}
read the original abstract

This article presents an automated method to quantify and detect symmetry elements in 2D patterns by means of image processing. Escher's woodcuts, a widely recognized didactic tool for crystallographic education of students, were used to demonstrate this approach. We also discuss peculiarities in the detection of black and white symmetry, color symmetry, and detection of the "hidden" and "broken" symmetry elements by means of the phase origin map approach.

Figures

Figures reproduced from arXiv: 1908.07932 by the authors.

Figure 1
Figure 1. (a) "Angel-Devil" (No.45), M.C.Escher, 1941, with a [PITH_FULL_IMAGE:figures/full_fig_p006_1.png] view at source ↗
Figure 2
Figure 2. (a)"Impossible figure" a hexagonal pattern with a [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗
Figure 3
Figure 3. (Program Screen shot) Figure 2a refined in the [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗
Figures from the paper (9 more)
Figure 4
Figure 4. Figure 4: POM for the p31m group (apparent) high symmetry to lower by exclusion of some symmetry operators due to coloring, defects, twinning, grain boundaries, magnetization, etc. 4.3 Example 3:"Lizards" This example is devoted to the problem of detection and quantification of …
Figure 5
Figure 5. Figure 5: "Lizards" (No.124), M. C. Escher, 1965. gy-lines are shown in red, g 0 x - in blue, the unit cell of the b/w group p2 0 g 0 g is shown as well. Initial image processing is done the same way as before. Because of use of the autocorrelation during the peak search routine…
Figure 6
Figure 6. Figure 6: (Program Screen shot) Figure 5 refined in the [PITH_FULL_IMAGE:figures/full_fig_p010_6.png]
Figure 7
Figure 7. Figure 7: (a) "Gray image" refined in p2gg group. (b) Lizards silhouettes (colorblind image) Figure 7b). While there is a slight difference in the lizards’ eyes in Figure 7b, all eyes are averaged out after refinement in the p2gg group. See [PITH_FULL_IMAGE:figures/full_fig_p01…
Figure 8
Figure 8. Figure 8: (Program Screen shot) Figure 7b refined in [PITH_FULL_IMAGE:figures/full_fig_p011_8.png]
Figure 9
Figure 9. Figure 9: "Reptiles", M. C. Escher, 1943, Fragment, with a [PITH_FULL_IMAGE:figures/full_fig_p011_9.png]
Figure 10
Figure 10. Figure 10: (a) Indexed Fourier Transformation of Figure 9. (b) POM for [PITH_FULL_IMAGE:figures/full_fig_p012_10.png]
Figure 11
Figure 11. Figure 11: Symmetrization of a uniformed silhouette image of Reptiles [PITH_FULL_IMAGE:figures/full_fig_p013_11.png]
Figure 12
Figure 12. Figure 12: (a) Schematic representation of a set of symmetry related FCs. (b) Graphical representation of the [PITH_FULL_IMAGE:figures/full_fig_p014_12.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

26 extracted references · 26 canonical work pages

  1. [1]

    Schwarzenbach

    D. Schwarzenbach. Crystallography. John Wiley, 1996

  2. [2]

    Gould and W

    R.O. Gould and W. Massa. Crystal Structure Determination. Springer Berlin Heidelberg, 2013

  3. [3]

    Shubnikov and V .A

    A.V . Shubnikov and V .A. Koptsik.Symmetry in Science and Art . Plenum Press, 1974

  4. [4]

    Shubnikov and N.V

    A.V . Shubnikov and N.V . Belov.Colored Symmetry. Macmillan, 1964

  5. [5]

    The mystery of the fifteenth bravais lattice

    Allen Nussbaum. The mystery of the fifteenth bravais lattice. American Journal of Physics , 68(10):950–954, 2000

  6. [6]

    Amezcua-López and A

    J. Amezcua-López and A. E. Cordero-Borboa. A mechanical lattice aid for crystallography teaching. The Physics Teacher, 26(6):384–389, 1988

  7. [7]

    Visual discrimination of the 17 plane symmetry groups

    Klaus Landwehr. Visual discrimination of the 17 plane symmetry groups. Symmetry, 3(2):207 – 219, 01 2011

  8. [8]

    L. Koren. Wabi-sabi for Artists, Designers, Poets & Philosophers . Wabi-sabi for Artists, Designers, Poets & Philosophers. Imperfect Publishing, 2008

Show all 26 references
  1. [9]

    Structure Determination from HREM by Crystallographic Image Processing , pages 275–300

    Xiaodong Zou and Sven Hovmöller. Structure Determination from HREM by Crystallographic Image Processing , pages 275–300. Springer Netherlands, Dordrecht, 2006

  2. [10]

    From virus structure to chromatin: X-ray diffraction to three-dimensional electron microscopy

    Aaron Klug. From virus structure to chromatin: X-ray diffraction to three-dimensional electron microscopy. Annual Review of Biochemistry, 79(1):1–35, 2010. PMID: 20192760

  3. [11]

    Electron crystallography: Electron microscopy and electron diffraction

    Xiaodong Zou, Sven Hovmöller, and Peter Oleynikov. Electron crystallography: Electron microscopy and electron diffraction. Electron Crystallography: Electron Microscopy and Electron Diffraction, pages 1–344, 01 2012

  4. [12]

    P. Moeck. Systems for assessing and enhancing the performance of scanning probe microscopes by quantifying and enforcing symmetries and periodicities in two dimensions, jun " 5" 2012. US Patent 8,196,216

  5. [13]

    Straton, Bill Moon, Taylor T

    Jack C. Straton, Bill Moon, Taylor T. Bilyeu, and Peter Moeck. Removal of multiple-tip artifacts from scanning tunneling microscope images by crystallographic averaging. Advanced Structural and Chemical Imaging , 1(1):14, Nov 2015

  6. [14]

    International Tables for Crystallography, V olume A: Space Group Symmetry

    Theo Hahn, editor. International Tables for Crystallography, V olume A: Space Group Symmetry. International Tables for Crystallography. Kluwer Academic Publishers, Dordrecht, Boston, London, 5. revised edition edition, 2002

  7. [15]

    Shmueli, editor

    U. Shmueli, editor. International Tables for Crystallography, V olume B: Reciprocal Space. International Tables for Crystallography. Springer Netherlands, 2008

  8. [16]

    X. Zou. Electron Crystallography of Inorganic Structures: Theory and Practice . Chemical Communications. Stockholm University, 1997

  9. [17]

    Symmetry-adapted fourier series for the wallpaper groups

    Bart Verberck. Symmetry-adapted fourier series for the wallpaper groups. Symmetry, 4(3):379–426, 2012

  10. [18]

    A. V . Oppenheim and J. S. Lim. The importance of phase in signals.Proceedings of the IEEE, 69(5):529–541, May 1981

  11. [19]

    Crisp: crystallographic image processing on a personal computer

    Sven Hovmöller. Crisp: crystallographic image processing on a personal computer. Ultramicroscopy, 41(1):121 – 135, 1992

  12. [20]

    Crystallographic processing of scanning probe microscopy images on the example of 2d periodic molecular monolayer arrays

    P Moeck, J Straton, B Moon, P Plachinda, M Hietschold, H Glowatzki, N Koch, and JP Rabe. Crystallographic processing of scanning probe microscopy images on the example of 2d periodic molecular monolayer arrays. Microscopy and Microanalysis, 16(S2):456–457, 2010

  13. [21]

    Moeck, P

    P. Moeck, P. Plachinda, B. Moon, J. Straton, S. Rouvimov, M. Toader, M. Abdel-Hafiez, and M. Hietschold. Quantifying and enforcing two-dimensional symmetries in scanning probe microscopy images of periodic objects. In 2009 International Semiconductor Device Research Symposium ,...

  14. [22]

    Crystallographic image processing software for scanning probe microscopists

    Pavel Plachinda, Bill Moon, and Peter Moeck. Crystallographic image processing software for scanning probe microscopists. APS March Meeting, 2010

  15. [23]

    Automatic lattice determination for two-dimensional crystal images

    Xiangyan Zeng, Bryant Gipson, Zi Yan Zheng, Ludovic Renault, and Henning Stahlberg. Automatic lattice determination for two-dimensional crystal images. Journal of Structural Biology, 160(3):353 – 361, 2007. Electron Crystallography of Membrane Proteins

  16. [24]

    Towards generalized noise-level dependent crystallographic symmetry classifications of more or less periodic crystal patterns

    Peter Moeck. Towards generalized noise-level dependent crystallographic symmetry classifications of more or less periodic crystal patterns. Symmetry, 10(5), 2018

  17. [25]

    Gramlich and H

    V . Gramlich and H. Wondratschek.Graphs for klassengleiche subgroups, pages 415–425. Springer Netherlands, Dordrecht, 2004

  18. [26]

    Selection is Distortion, pages 5–16

    Bruno Ernst and Hans de Rijk. Selection is Distortion, pages 5–16. Springer Berlin Heidelberg, 2007. 16

Pith tools

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