The DOI matches its Crossref record; 91 of the 91 references it cites were found in a registry or an index; 11 of the 72 sources its claims cite were checked for support.
Topological pattern recognition for point cloud data
Open accessAbstract
In this paper we discuss the adaptation of the methods of homology from algebraic topology to the problem of pattern recognition in point cloud data sets. The method is referred to aspersistent homology, and has numerous applications to scientific problems. We discuss the definition and computation of homology in the standard setting of simplicial complexes and topological spaces, then show how one can obtain useful signatures, called barcodes, from finite metric spaces, thought of as sampled from a continuous object. We present several different cases where persistent homology is used, to illustrate the different ways in which the method can be applied.
Overview
CoChat Summary
TL;DR
This paper reviews persistent homology, adapting algebraic topology’s homology to finite metric spaces and point cloud data, and presents barcodes/persistence diagrams as signatures. It surveys theory, stability, and applications including natural image patches, compound databases, viral evolution, time series, and the cosmic web.
Key findings
- Persistent homology replaces Betti numbers with barcodes/persistence diagrams; long intervals correspond to larger-scale geometric features, while short intervals are typically treated as noisy.
- Stability results: the map β_k from finite metric spaces under Gromov-Hausdorff distance to barcodes under bottleneck distance is distance non-increasing (Theorem 4.2), and functional persistence has a stability theorem for tame functions (Theorem 4.3).
- Natural image patches: 4×10^6 high-contrast 3×3 patches (top 20% D-norm) lie on a seven-dimensional ellipsoid E7 in an eight-dimensional subspace of R9; X(300,30) shows one long 1D bar suggesting a circle, X(15,30) shows five long bars suggesting β1=5, and the full data suggests a Klein bottle, with mod 3 Betti numbers distinguishing torus (1,2,1) from Klein bottle (1,1,0).
- Applications: a 4000-compound DFHR inhibitor database localized into four groups via Mapper; viral evolution uses homology to detect reassortment/recombination; sliding-window persistence detects periodicity; cosmic web Betti numbers carry more information than the Euler characteristic alone.
Methods
- Design: expository review of algebraic topology and persistent homology, with definitions, theorems, and worked examples rather than a primary empirical study.
- Data/sample: illustrative datasets include 4×10^6 high-contrast 3×3 image patches (top 20% D-norm), random subsamples of 5×10^4, a 4000-compound database, viral sequences, time series, and cosmic web data.
- Analysis: construct Vietoris-Rips, alpha, witness, and functional persistence complexes; compute homology over fields via matrix row/column operations; represent invariants as barcodes/persistence diagrams; compare with bottleneck/Wasserstein and Gromov-Hausdorff metrics; extensions include zigzag and multidimensional persistence.
Limitations
- Paper-stated: persistent homology computation has the same complexity as Gaussian elimination, n^3 for an n×n matrix, so direct calculation is expensive; Vietoris-Rips complexes are often too large; alpha complexes become intractable as embedding dimension N increases; witness complex accuracy is difficult to assess.
- Paper-stated: there is no universal or natural choice of scale ρ; multidimensional persistence lacks a complete barcode-like classification; the barcode space (B∞, d_p) is not complete, and Fréchet means may be empty or non-unique.
- Evident from text: as a review, application results are summarized from cited studies rather than newly validated here.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
10 Valid · 10 Not valid · 53 Not checked · 155 Own results and interpretations
Every source the paper cites for this claim was checked and supports it.
In the paper
“One obstacle to realizing this goal is the fact that the metric space ( B ∞ , d p ) is not complete.”
What the sources say
- Supportedcites Mileyko et al. (2011) source
In §2.4, following Definition 1 (page unavailable), the paper says that the space of persistence diagrams under the pth Wasserstein distance is not complete and gives a Cauchy sequence with no limit in that space. It then extends the definition to generalized diagrams; its later completeness result concerns that enlarged space. (LLM, conf=0.88)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“if we restrict the domain of the scale parameter r in ordinary persistence to the integer lattice Z + ⊆ R + , then the classification of such restricted persistence vector spaces is identical to the classification of graded k [ t ]-modules.”
What the sources say
- Partly supportedcites Zomorodian and Carlsson (2005) sourceExperimental — accuracy not yet measured
Theorem 3.1 (Section 3.1, p. 11) establishes an equivalence between integer-indexed persistence modules and graded R[t]-modules, supporting the claimed correspondence when R is the field k. But the theorem specifies persistence modules of finite type and finitely generated, non-negatively graded modules; the claim omits those qualifications. (LLM, conf=0.85)
plainstates an association- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“More recently, Park et al. (2013) and van de Weygaert et al. (2011) have shown that the Betti numbers of Gaussian random fields carry strictly more information concerning the random field than the Euler characteristic alone.”
What the sources say
- Supportedcites Park et al. (2013) source
The Introduction (Sec. 1) states that "the topology of figures can be specified by the Betti numbers and the genus of the surface of the figures is only a linear combination of them. Therefore, Betti numbers are a more versatile and elaborate characterization of topology," and Sec. 2 makes this explicit for Gaussian excursion sets with g ≡ -χ(∂M)/2 = -χ(M) = -β0 + β1 - β2 via the Euler-Poincaré theorem. The paper then backs this with numerics on Gaussian random fields (Sec. 3, Sec. 5), showing the three Betti curves each track a different extremum of the genus curve and depend on the spectral index differently, so the single genus/Euler-characteristic value does not recover them. The claim's "strictly more information than the Euler characteristic alone" is thus a fair paraphrase of what our copy says for Gaussian random fields; the co-citation (van de Weygaert et al. 2011) covers only the other half of the citing sentence. (LLM, conf=0.85)
- Supportedcites van de Weygaert et al. (2011) source
Section 8.6 'Genus and Homology' (p. 33-34) states for Gaussian random fields that the genus function's shape is independent of the power spectrum while the Betti numbers do reflect it, with substantial differences between fields of different power spectra, making them 'potentially strong discriminators between cosmological scenarios' — i.e. strictly more information than the genus/Euler characteristic. Section 8.7 (p. 34) adds that the Betti numbers in Gaussian fields are not independent of one another yet 'contains complementary information on the topological structure', and Section 2.1 (p. 8-9) notes the Euler characteristic is only the alternating sum of the Betti numbers, so it cannot recover them individually. The Abstract (p. 1) frames the whole paper as extending topological information 'beyond conventional cosmological studies of topology in terms of genus and Euler characteristic'. (LLM, conf=0.82)
boostedstates an association- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Rests on an assumption not shown here.
- Supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“In the cases studied, these occurrences were consistent with the known horizontal transfer mechanisms in these situations.”
What the sources say
- Supportedcites Chan et al. (2013) source
Our copy repeatedly reports that the topological features (cycles/polytopes) it detected matched the horizontal transfer mechanisms already known for each case: the H7N9 2D irreducible cycle 'recapitulates the triple reassortment' established by Gao et al. (Vertical Evolution in Influenza), the HIV b2 polytope 'identifies CRF13_cpx and CRF02_AE, two complex circulating recombinant forms', the avian PB2-PB1-PA-NP cosegregation pattern 'mimics previous in vitro results' on polymerase–NP interaction constraining reassortment, and the ICR estimates are 'supported by previous phylogenetic estimates' and 'correlate with the high proportion of avian reassortants reported in previous studies'. No single sentence states the blanket generalization verbatim, but the paper treats the topic that way across the cases studied, so the paraphrase is supported; the only mild caveat is that the H1N1pdm cycles nominate candidate reassortments rather than confirm a previously documented event. (LLM, conf=0.80)
plainstates an association- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“We begin with the first three letters of the alphabet, and find that H 1 succeeds in distinguishing between them.”
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
106 more claims
Create a free account to see all 228 claims and to verify the 27 references still open.
- It may state two separate claims as one. 52
- Probably not a substantial claim. 81
- The quote only partly states this claim. 1
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
Each contradiction between two passages listed here was confirmed by an independent check.
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ 2.7. Making homology more sensitive · p. 23
“| β 1 = 0, β 0 ( e ( X )) = 1 | none |”
§ 2.7. Making homology more sensitive · p. 23
“| β 1 = 0, β 0 ( e ( X )) = 1 | { P , Q } |”
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ 5.1. Natural image patches · p. 57
“For a data set X , we let X ( k, p ), where k is a positive integer and p is a percentage, denote the set of points x for which the value of ρ k ( x ) is among the top p % of the values occurring in X .”
§ 5.1. Natural image patches · p. 57
“What it demonstrates is that the highest-density patches, according to ρ 300 , consist of the discrete versions of linear functions in two variables.”
Citations
Cites
Categories for the Working Mathematician
Lane, S. M. (1971) · Graduate Texts in Mathematics · doi:10.1007/978-1-4612-9839-7 · Cited by 8,817
Bhatia, R., Bredon, G. E., Walter, W. L., Rotman, J. J., Murty, M. R., Gilman, J., Walters, P., Golubitsky, M., Karatzas, I., Cohen, H., et al. (1977) · Cited by 8,252
Three‐Year Wilkinson Microwave Anisotropy Probe ( WMAP ) Observations: Implications for Cosmology
Spergel, D. N., Bean, R., Dore, O., Nolta, M. R., Bennett, C. L., Dunkley, J., Hinshaw, G., Jarosik, N., Komatsu, E., Page, L., et al. (2007) · The Astrophysical Journal Supplement Series · doi:10.1086/513700 · Cited by 6,778
(1975) · CERN Document Server (European Organization for Nuclear Research) · Cited by 5,958
(2001) · Cited by 5,374
Lint, J. H. (2020) · Cited by 5,023
ZINC − A Free Database of Commercially Available Compounds for Virtual Screening
Irwin, J. J., & Shoichet, B. K. (2005) · Journal of Chemical Information and Modeling · doi:10.1021/ci049714+ · Cited by 4,058
The statistics of peaks of Gaussian random fields
Bardeen, J. M., Bond, J. R., Kaiser, N., & Szalay, A. S. (1986) · The Astrophysical Journal · doi:10.1086/164143 · Cited by 3,270
Wooding, S. (2004) · The American Journal of Human Genetics · doi:10.1086/383584 · Cited by 3,032
Burago, D., Burago, Y., & Ivanov, S. (2001) · Graduate Studies in Mathematics · doi:10.1090/gsm/033 · Cited by 2,598
Research works printed but not linked: 8
- Fréchet, 1944, L'intégrale abstraite d'une fonction abstraite d'une variable abstraite et son application à la moyenne d'un élément aléatoire de nature quelconque, Rev. Sci., 82, 483
- Euler, 1758, Demonstratio nonnullarum insignium proprietatum quibus solida hedris planis inclusa sunt praedita, Novi Commentarii Academiae Scien-tiarum Petropolitanae, 4, 140
- Dalbec, 1999, Multisymmetric functions, Beiträge Algebra Geom., 40, 27
- de Silva, 2004, Proc. First Eurographics Conference on Point-Based Graphics, 157
- Singh, 2007, Proc. Eurographics Symposium on Point-Based Graphics 2007
- Zomorodian, 2005, Cambridge Monographs on Applied and Computational Mathematics, 16
- Robins, 1999, Towards computing homology from finite approximations, Topology Proceedings, 24, 503
- Vandermonde, 1774, Mémoires de l'Académie Royale des Sciences pour 1771, 556
Cited by
Virtual, Digital and Hybrid Twins: A New Paradigm in Data-Based Engineering and Engineered Data
Chinesta, F., Cueto, E., Abisset-Chavanne, E., Duval, J. L., & Khaldi, F. E. (2020) · Archives of Computational Methods in Engineering · doi:10.1007/s11831-018-9301-4 · Cited by 299
Cang, Z., Mu, L., & Wei, G.-W. (2018) · PLOS Computational Biology · doi:10.1371/journal.pcbi.1005929 · Cited by 257
Towards a new approach to reveal dynamical organization of the brain using topological data analysis
Saggar, M., Sporns, O., Gonzalez-Castillo, J., Bandettini, P. A., Carlsson, G., Glover, G., & Reiss, A. L. (2018) · Nature Communications · doi:10.1038/s41467-018-03664-4 · Cited by 253
Self-citationLim, L.-H. (2020) · SIAM Review · doi:10.1137/18m1223101 · Cited by 234
“Topology can be a useful tool in such contexts [17] since it requires very little — essentially just a weak notion of separation, i.”
MethodThe Theory of the Interleaving Distance on Multidimensional Persistence Modules
Lesnick, M. (2015) · Foundations of Computational Mathematics · doi:10.1007/s10208-015-9255-y · Cited by 203
“In the last decade and a half, it has been applied widely [9,32] and has been the subject of a large body of theoretical work.”
BackgroundInfluential
Metadata
- Authors
- Carlsson, G.
- Affiliations
- Stanford University (US)
- Corresponding author
- Carlsson, Gunnar
- ORCID
- Carlsson, Gunnar
- Venue
- Acta Numerica
- Publisher
- Cambridge University Press (CUP)
- Published
- 2014-05
- Volume
- 23
- Pages
- 289-368
- ISSN
- 0962-4929 (ISSN-L)
- 1474-0508
- OpenAlex
- W2148676644
- Cited by
- 230
- References
- 88
- Citations by year
- 2026: 7 · 2025: 12 · 2024: 15 · 2023: 28 · 2022: 13 · 2021: 24 · 2020: 27 · 2019: 30 · 2018: 31 · 2017: 14 · 2016: 17 · 2015: 10 · 2014: 2
- FWCI
- 14.80
- Type
- journal article
- Language
- en
- Open access
- bronze
- doi.org
- Locations
- Acta Numerica (publishedVersion)
- Funders
- National Science Foundation
- University of Cambridge
- National Institutes of Health
- Air Force Office of Scientific Research
- Topics
- Physical Sciences › Computer Science › Computational Theory and Mathematics › Topological and Geometric Data Analysis
- Physical Sciences › Mathematics › Mathematical Physics › Homotopy and Cohomology in Algebraic Topology
- Physical Sciences › Mathematics › Geometry and Topology › Geometric and Algebraic Topology
- Keywords
- persistent homology, point cloud data, pattern recognition, simplicial complexes, barcodes, finite metric spaces, topological signatures, algebraic topology
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
The claim verdicts this report counts are experimental: their accuracy is not yet measured.
A language model drew the claims from the paper on 2026-10-10, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-10-10 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Cited sources are checked with a language model’s help (gpt-6-sol and qwen3.8-omni-flash).
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
How to cite
Cite CoChat for the checks on this page, and the paper for its own findings.
- This page
- CoChat. (2026). Citation check: Topological pattern recognition for point cloud data. https://app.cochat.ai/papers/10.1017/s0962492914000051
- The paper
- Carlsson, G. (2014). Topological pattern recognition for point cloud data. Acta Numerica. https://doi.org/10.1017/s0962492914000051
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.