Pith. sign in

REVIEW 3 major objections 4 minor 35 references

Morphological classification of eclipsing binary stars using computer vision methods

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

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

Pith's one-line read The paper claims that computer vision models fine-tuned on synthetic polar-hexbin light-curve images classify eclipsing binaries as detached or overcontact with over 96% accuracy on validation data, remain above 94% (up to 100%) on external

desk verdict A genuinely new image representation for phase-folded light curves, with honest reporting of its own spot-detection failure; the headline binary accuracy is plausible but the abstract leaves too little detail to take the transfer claim to the bank. read the letter →

arxiv 2508.12802 v1 pith:J6H3FVE7 submitted 2025-08-18 cs.CV astro-ph.IMastro-ph.SR

Štefan Parimucha , Maksim Gabdeev , Yanna Markus , Martin Vaňko , Pavol Gajdoš This is my paper
classification cs.CVastro-ph.IMastro-ph.SR
keywords eclipsingbinarieslightcurveclassificationcomputervisionconvolutionalneuralnetworkstransformerssynthetictrainingdatastarspotdetectionpolarhexbinrepresentation
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

This paper asks whether a generic image-classification pipeline can sort eclipsing binary stars into morphological classes directly from light-curve images. The authors fine-tune a convolutional network (ResNet50) and a vision transformer on synthetic light curves that have been phase-folded, mapped into polar coordinates, and rendered as hexbin images. On validation data the detached-versus-overcontact classifier exceeds 96% accuracy across the Gaia G, I, and TESS passbands, and on external catalogues (OGLE, DEBCat, WUMaCat) it stays above 94%, reaching 100% for TESS. The second, finer-grained task—automated spot detection—performs poorly, and the paper identifies this as a clear limitation. The overall message is that coarse morphological classification by computer vision can scale to large surveys, while subtle photometric features still need a different approach.

What carries the argument

The central object is the polar-coordinate hexbin image representation: each eclipsing-binary light curve is phase-folded, mapped into polar coordinates, and binned into a hexbin grid to form an image that pre-trained CNNs and vision transformers can ingest. This representation, together with fine-tuning on synthetic light curves and a hierarchical two-stage classifier (detached/overcontact first, spots second), carries the whole argument: the paper's accuracy numbers are demonstrated on images built this way, and the spot-detection failure is likewise attributed to what the representation and synthetic training do and do not preserve.

What would settle it

Take an independent set of real eclipsing binaries whose detached or overcontact status is already established by non-photometric means (e.g., spectroscopic radial velocities and derived Roche geometry) and run the trained classifier on their survey light curves. If accuracy on this set drops far below the 94–100% range reported on the paper's catalogues, the synthetic-to-real transfer is the weak link. A second falsifier: add synthetic images with realistic spot signals and instrument noise and check whether spot-detection accuracy rises; if it does, the current spot failure is a training-dat

Watch

Extended reading notes

Core claim

The paper's central claim is that generic image-classification models, fine-tuned on synthetic light-curve images, can classify real eclipsing binaries by coarse morphology. The authors map each phase-folded light curve into polar coordinates and render it as a hexbin image, then fine-tune ResNet50 and a vision transformer in a two-stage hierarchy: detached versus overcontact first, spot presence second. The first stage exceeds 96% validation accuracy across Gaia G, I, and TESS passbands and 94–100% on the OGLE, DEBCat, and WUMaCat catalogues. The second stage, spot detection, performs poorly, a limitation the paper states explicitly. Together these results say the method transfers for broad

Load-bearing premise

The load-bearing premise is that synthetic light-curve images are realistic enough that a model fine-tuned on them recognizes real survey light curves the same way; the paper's own poor spot-detection results show this transfer is incomplete, so the high morphology accuracy could partly reflect properties of the synthetic images or label definitions rather than fully generalizable astrophysical morphology.

Editorial extensions

If this is right

  • Large photometric surveys can use the same two-class image pipeline as a fast pre-screen for detached versus overcontact eclipsing binaries, reducing the volume needing full physical modeling.
  • The method's multi-passband success (Gaia G, I, and TESS) suggests the learned morphology signal is not tied to one instrument's noise properties.
  • The hierarchical structure means a failed second stage does not necessarily degrade the first: morphology classification can proceed even where spot detection cannot.
  • The spot-detection failure implies that distinguishing subtle features likely requires either more realistic synthetic training images, dedicated feature engineering, or different supervision than coarse morphology.
  • External catalogue performance above 94% indicates the classifier generalizes beyond the synthetic training distribution at least for the broad detached/overcontact split.

Reading between the lines

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

  • An implicit corollary of the spot-detection failure is that the model's success on morphology likely rests on global, coarse shape features rather than on the pixel-level detail that would reveal spots; masking the out-of-eclipse regions in a controlled experiment would test this directly.
  • The polar-hexbin representation is not tied to eclipsing binaries; the same fine-tuning recipe could be tested on other phase-folded periodic variables, where the classification problem has the same image-morphism structure.
  • A stronger external check than the catalogue labels used here would be to compare predictions against an independent sample with spectroscopically determined binary geometry, which would reveal whether the synthetic-to-real transfer holds where the truth is known at the individual-object level.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 4 minor

Summary. The paper proposes a computer-vision pipeline for morphological classification of eclipsing binaries (EBs). Pre-trained ResNet50 and ViT models are fine-tuned on synthetic light curves transformed into a polar-coordinate hexbin image representation. The pipeline first classifies systems as detached versus overcontact, then attempts to detect spots. The abstract reports validation accuracies above 96% across Gaia G, I, and TESS passbands, and external test accuracies above 94% (up to 100% for TESS) on OGLE, DEBCat, and WUMaCat. The abstract also explicitly states that the spot-detection task performed poorly, a limitation the authors acknowledge.

Significance. If the reported external performance holds, this would be a valuable demonstration that modern computer vision models can classify EB morphology at scale from phase-folded light curves, with direct applicability to large photometric surveys. The explicit reporting of the spot-detection failure is a notable strength, as it avoids overclaiming. However, the abstract alone does not establish the central claim of robust transfer from synthetic training data to real survey data; the missing sample sizes, stratification, and methodological detail leave the generalization claim under-supported. The significance is therefore conditional on the availability of the full methods and a more granular evaluation.

major comments (3)
  1. [Abstract (external performance claims)] The abstract states external test performance of ">94%, up to 100% for TESS" on OGLE, DEBCat, and WUMaCat, but gives no sample sizes, class balance, or performance stratified by signal-to-noise, amplitude, orbital phase coverage, or passband. If these external catalogs are dominated by high-amplitude, well-sampled, high-SNR light curves, the accuracies could reflect trivial shape cues rather than robust morphology. This is load-bearing for the paper's practical claim, and the missing stratification must be supplied before the generalization claim can be accepted.
  2. [Abstract (spot-detection limitation)] The abstract reports that automated spot detection 'performed poorly,' which is direct evidence that the synthetic-to-real transfer is incomplete for subtle photometric features. Since both the binary classification and the spot-detection models are fine-tuned on the same synthetic distribution, this failure raises the risk that the binary classification also relies on synthetic artifacts, even if gross morphology masks it. The paper needs to justify why the binary transfer succeeds despite the spot-detection transfer failure, e.g., via feature-attribution analysis, ablation of synthetic artifacts, or failure analysis on misclassified external objects.
  3. [Abstract (polar-hexbin representation)] The polar hexbin representation may inadvertently encode sampling density, cadence, and noise properties of the light curve rather than astrophysical morphology. The abstract gives no details on how phase coverage, binning, or noise are normalized, so it is unclear whether the high external accuracies are robust to the diverse cadence and sampling of real surveys. A concrete test would be to evaluate the trained models on artificially subsampled or noise-injected versions of the external light curves; the paper should report such an analysis or describe how the representation controls for these factors.
minor comments (4)
  1. [Abstract (presentation)] The abstract uses '>96%' and '>94%' without confidence intervals or exact numbers; reporting exact accuracies with uncertainties would be more informative.
  2. [Abstract (terminology)] The passband notation 'Gaia G, I, and TESS' is ambiguous; 'I' presumably means an infrared passband, but it should be defined (e.g., Gaia RP, or an I-band filter) and the exact photometric band used for each catalog should be specified.
  3. [Abstract (catalog references)] WUMaCat is introduced without a reference; all external catalogs should be cited in the text.
  4. [General] The abstract should mention whether code and trained models are available for reproducibility, and specify the exact number of synthetic and real objects used.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity detected: the transfer claim is independently testable against external catalog labels.

full rationale

The paper's primary claim is empirical: a CNN/vision-transformer classifier fine-tuned on synthetic light-curve images is validated and then tested on real observational data from OGLE, DEBCat, and WUMaCat. The validation and test labels come from external catalog classifications, not from the model's own outputs or from the synthetic data generator. The synthetic training images carry known morphological labels, but the model must still learn a mapping from pixel features to those labels; there is no equation, fitted parameter, or self-citation that makes the predicted class equal to the input by construction. The acknowledged failure of spot detection is a limitation indicating a synthetic-to-real transfer gap, but it is not a circular step. No load-bearing self-citation, imported uniqueness theorem, or ansatz-smuggling via citation is evident from the provided text. The concern that synthetic training data may encode the same definitions as the external catalogs is a representativeness/transfer issue, not a circularity reduction, because the external labels are independent and the model is not tuned on those labels. Therefore the derivation chain is self-contained with respect to circularity.

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

The central result rests on three unstated domain assumptions: realism of synthetic training data, correctness of external catalog labels, and information preservation by the phase-folded polar-hexbin representation. No new physical entities are introduced.

free parameters (3)
  • Polar angle mapping of orbital phase
    The phase coordinate is mapped to polar angle; the exact scaling or offset is a design choice that changes the image geometry and thus what the CNN sees.
  • Hexbin grid resolution
    Number of hexagonal bins and their size are chosen by hand; it controls how light-curve points are aggregated into pixel intensities.
  • Training hyperparameters
    Learning rate, batch size, number of epochs, and augmentation schedule are not given in the abstract; these affect validation accuracy and overfitting.
assumptions (3)
  • domain assumption Synthetic light curves are representative of real eclipsing binary observations.
    The fine-tuning relies on the generative model that produced the synthetic images to capture the same morphology, noise, and passband effects found in real survey data.
  • domain assumption Catalog labels in OGLE, DEBCat, and WUMaCat are correct.
    External test accuracy is measured against these labels, so label errors directly cap reported accuracy.
  • domain assumption Phase-folding and normalization preserve morphological information.
    The classifier sees only phase-folded, normalized brightness images; any morphology that is lost in folding cannot be recovered.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Morphological classification of eclipsing binary stars using computer vision methods." pith.science (2026). https://pith.science/paper/J6H3FVE7

@misc{pith2026250812802,
  author       = {Pith},
  title        = {Pith review of: Morphological classification of eclipsing binary stars using computer vision methods},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/J6H3FVE7}},
  note         = {Machine review of arXiv:2508.12802}
}
abstract

We present an application of computer vision methods to classify the light curves of eclipsing binaries (EB). We have used pre-trained models based on convolutional neural networks ($\textit{ResNet50}$) and vision transformers ($\textit{vit\_base\_patch16\_224}$), which were fine-tuned on images created from synthetic datasets. To improve model generalisation and reduce overfitting, we developed a novel image representation by transforming phase-folded light curves into polar coordinates combined with hexbin visualisation. Our hierarchical approach in the first stage classifies systems into detached and overcontact types, and in the second stage identifies the presence or absence of spots. The binary classification models achieved high accuracy ($>96\%$) on validation data across multiple passbands (Gaia~$G$, $I$, and $TESS$) and demonstrated strong performance ($>94\%$, up to $100\%$ for $TESS$) when tested on extensive observational data from the OGLE, DEBCat, and WUMaCat catalogues. While the primary binary classification was highly successful, the secondary task of automated spot detection performed poorly, revealing a significant limitation of our models for identifying subtle photometric features. This study highlights the potential of computer vision for EB morphological classification in large-scale surveys, but underscores the need for further research into robust, automated spot detection.

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

35 extracted references · 17 canonical work pages

  1. [1]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION word.in bbl.in ":" * " " * FUNCTION f...

  2. [2]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION word.in bbl.in ":" * " " * FUNCTION f...

  3. [3]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION word.in bbl.in ":" * " " * FUNCTION f...

  4. [4]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION word.in bbl.in ":" * " " * FUNCTION f...

  5. [5]

    Deep multi-survey classification of variable stars

    author Aguirre , C. , author Pichara , K. , author Becker , I. , year 2019 . title Deep multi-survey classification of variable stars . journal volume 482 , pages 5078--5092 . :10.1093/mnras/sty2836, http://arxiv.org/abs/1810.09440 arXiv:1810.09440

    work page Pith review arXiv 2019

  6. [6]

    , author Littlefield, R.J

    author Carr, D.B. , author Littlefield, R.J. , author Nicholson, W.L. , author Littlefield, J.S. , year 1987 . title Scatterplot matrix techniques for large n . journal Journal of the American Statistical Association volume 82 , pages 424--436 . http://www.jstor.org/stable/2289444

    work page arXiv 1987

  7. [7]

    , author Kurucz , R.L

    author Castelli , F. , author Kurucz , R.L. , year 2003 . title New Grids of ATLAS9 Model Atmospheres , in: editor Piskunov , N. , editor Weiss , W.W. , editor Gray , D.F. (Eds.), booktitle Modelling of Stellar Atmospheres , p. pages A20 . :10.48550/arXiv.astro-ph/0405087, http://arxiv.org/abs/astro-ph/0405087 arXiv:astro-ph/0405087

    arXiv doi:10.48550/arxiv.astro-ph/0405087 2003

  8. [8]

    , author Beyer, L

    author Dosovitskiy, A. , author Beyer, L. , author Kolesnikov, A. , author Weissenborn, D. , author Zhai, X. , author Unterthiner, T. , author Dehghani, M. , author Minderer, M. , author Heigold, G. , author Gelly, S. , author Uszkoreit, J. , author Houlsby, N. , year 2021 . title An image is worth 16x16 words: Transformers for image recognition at scale ...

    arXiv 2021

Show all 35 references
  1. [9]

    , author Zhang, X

    author He, K. , author Zhang, X. , author Ren, S. , author Sun, J. , year 2016 . title Deep residual learning for image recognition , in: booktitle Proceedings of the IEEE conference on computer vision and pattern recognition , pp. pages 770--778

    2016

  2. [10]

    , year 2001

    author Hilditch , R.W. , year 2001 . title An Introduction to Close Binary Stars

    2001

  3. [11]

    , author Bailer-Jones , C.A.L

    author Kim , D.W. , author Bailer-Jones , C.A.L. , year 2016 . title A package for the automated classification of periodic variable stars . journal volume 587 , pages A18 . :10.1051/0004-6361/201527188, http://arxiv.org/abs/1512.01611 arXiv:1512.01611

    2016 arXiv

  4. [12]

    , year 1959

    author Kopal , Z. , year 1959 . title Close binary systems

    1959

  5. [13]

    , author Parimucha , S

    author Kudak , V. , author Parimucha , S . , author Perig , V. , author Gajdo s , P. , year 2023 . title Photometric Analysis of Eclipsing Binaries: VY UMi, RU UMi and GSC 04364-00648 . journal volume 59 , pages 137 . :10.22201/ia.01851101p.2023.59.01.10, http://arxiv.org/abs/...

    2023 arXiv

  6. [14]

    , author C eki , A

    author Latkovi \'c , O. , author C eki , A. , author Lazarevi \'c , S. , year 2021 . title Statistics of 700 Individually Studied W UMa Stars . journal volume 254 , pages 10 . :10.3847/1538-4365/abeb23, http://arxiv.org/abs/2103.06693 arXiv:2103.06693

    2021 arXiv

  7. [15]

    , author Horne , K

    author Marsh , T.R. , author Horne , K. , year 1988 . title Accretion disc mapping . journal Advances in Space Research volume 8 , pages 127--134 . :10.1016/0273-1177(88)90395-X

    1988 doi

  8. [16]

    , author Holl , B

    author Mowlavi , N. , author Holl , B. , author Lecoeur-Ta \" bi , I. , author Barblan , F. , author Kochoska , A. , author Pr s a , A. , author Mazeh , T. , author Rimoldini , L. , author Gavras , P. , author Audard , M. , author Jevardat de Fombelle , G. , author Nienartowic...

    2023 arXiv

  9. [17]

    , author Ranasinghe, K

    author Naseer, M. , author Ranasinghe, K. , author Khan, S.H. , author Hayat, M. , author Khan, F.S. , author Yang, M. , year 2021 . title Intriguing properties of vision transformers . journal CoRR volume abs/2105.10497 . https://arxiv.org/abs/2105.10497, http://arxiv.org/abs...

    2021 arXiv

  10. [18]

    , author Hauschildt , P.H

    author Orosz , J.A. , author Hauschildt , P.H. , year 2000 . title The use of the NextGen model atmospheres for cool giants in a light curve synthesis code . journal Astronomy and Astrophysics volume 364 , pages 265--281 . http://arxiv.org/abs/astro-ph/0010114 arXiv:astro-ph/0010114

    2000 arXiv

  11. [19]

    , author Gajdo s , P

    author Parimucha , S . , author Gajdo s , P. , author Markus , Y. , author Kudak , V. , year 2024 . title Deep-learning classification of eclipsing binaries . journal Contributions of the Astronomical Observatory Skalnate Pleso volume 54 , pages 167--170 . :10.31577/caosp.2024...

    2024 doi

  12. [20]

    , author Soszy \'n ski , I

    author Pawlak , M. , author Soszy \'n ski , I. , author Udalski , A. , author Szyma \'n ski , M.K. , author Wyrzykowski , . , author Ulaczyk , K. , author Poleski , R. , author Pietrukowicz , P. , author Koz owski , S. , author Skowron , D.M. , author Skowron , J. , author Mr ...

    2016 doi arXiv

  13. [21]

    , year 2007

    author Percy, J.R. , year 2007 . title Understanding variable stars . publisher Cambridge University Press

    2007

  14. [22]

    , author Chochol , D

    author Pribulla , T. , author Chochol , D. , author Milano , L. , author Errico , L. , author Vittone , A.A. , author Barone , F. , author Parimucha , S . , year 2000 . title Active eclipsing binary RT Andromedae revisited . journal volume 362 , pages 169--188

    2000

  15. [23]

    , author Conroy , K

    author Prsa , A. , author Conroy , K. , author Horvat , M.a.a. , year 2016 . title Physics of eclipsing binaries. ii. towards the increased model fidelity . journal The Astrophysical Journal Supplement Series volume 227

    2016

  16. [24]

    , year 2018

    author Prša, A. , year 2018 . title Modeling and Analysis of Eclipsing Binary Stars . 2514-3433, publisher IOP Publishing . https://dx.doi.org/10.1088/978-0-7503-1287-5, :10.1088/978-0-7503-1287-5

    2018 doi

  17. [25]

    , author Starr , D.L

    author Richards , J.W. , author Starr , D.L. , author Butler , N.R. , author Bloom , J.S. , author Brewer , J.M. , author Crellin-Quick , A. , author Higgins , J. , author Kennedy , R. , author Rischard , M. , year 2011 . title On Machine-learned Classification of Variable Sta...

    2011 arXiv

  18. [26]

    , author Winn , J.N

    author Ricker , G.R. , author Winn , J.N. , author Vanderspek , R. , author Latham , D.W. , author Bakos , G. \'A . , author Bean , J.L. , author Berta-Thompson , Z.K. , author Brown , T.M. , author Buchhave , L. , author Butler , N.R. , author Butler , R.P. , author Chaplin ,...

    2015

  19. [27]

    , author Holl , B

    author Rimoldini , L. , author Holl , B. , author Gavras , P. , author Audard , M. , author De Ridder , J. , author Mowlavi , N. , author Nienartowicz , K. , author Jevardat de Fombelle , G. , author Lecoeur-Ta \" bi , I. , author Karbevska , L. , author Evans , D.W. , author ...

    2023 arXiv

  20. [28]

    , author S \'a nchez-Fern \'a ndez , C

    author Sarro , L.M. , author S \'a nchez-Fern \'a ndez , C. , author Gim \'e nez , \'A . , year 2006 . title Automatic classification of eclipsing binaries light curves using neural networks . journal volume 446 , pages 395--402 . :10.1051/0004-6361:20052830, http://arxiv.org/...

    2006 arXiv

  21. [29]

    , author Chen , J

    author Shan , Y. , author Chen , J. , author Zhang , Z. , author Wang , L. , author Zou , Z. , author Li , M. , year 2025 . title Identifying Eclipsing Binary Stars with TESS Data Based on a New Hybrid Deep Learning Model . journal volume 137 , pages 044503 . :10.1088/1538-387...

    2025 arXiv

  22. [30]

    , author Pawlak , M

    author Soszy \'n ski , I. , author Pawlak , M. , author Pietrukowicz , P. , author Udalski , A. , author Szyma \'n ski , M.K. , author Wyrzykowski , . , author Ulaczyk , K. , author Poleski , R. , author Koz owski , S. , author Skowron , D.M. , author Skowron , J. , author Mr ...

    2016 doi arXiv

  23. [31]

    , year 2015

    author Southworth , J. , year 2015 . title DEBCat: A Catalog of Detached Eclipsing Binary Stars , in: editor Rucinski , S.M. , editor Torres , G. , editor Zejda , M. (Eds.), booktitle Living Together: Planets, Host Stars and Binaries , p. pages 164 . :10.48550/arXiv.1411.1219,...

    2015 doi arXiv

  24. [32]

    , author Fedurco , M

    author C okina , M. , author Fedurco , M. , author Parimucha , S . , year 2021 a. title ELISa: A new tool for fast modelling of eclipsing binaries . journal volume 652 , pages A156 . :10.1051/0004-6361/202039171, http://arxiv.org/abs/2106.10116 arXiv:2106.10116

    2021 arXiv

  25. [33]

    , author Maslej-Kre s n \'a kov \'a , V

    author C okina , M. , author Maslej-Kre s n \'a kov \'a , V. , author Butka , P. , author Parimucha , S . , year 2021 b. title Automatic classification of eclipsing binary stars using deep learning methods . journal Astronomy and Computing volume 36 , pages 100488 . :10.1016/j...

    2021

  26. [34]

    , author Bai , Y

    author Wang , C. , author Bai , Y. , author Han , H. , author Yang , H. , author Liu , J. , year 2023 . title Transfer Learning Applied to Stellar Light Curve Classification . journal arXiv e-prints , pages arXiv:2305.13745 :10.48550/arXiv.2305.13745, http://arxiv.org/abs/2305...

    2023 doi arXiv

  27. [35]

    , year 1979

    author Wilson , R.E. , year 1979 . title Eccentric orbit generalization and simultaneous solution of binary star light and velocity curves. journal volume 234 , pages 1054--1066 . :10.1086/157588

    1979 doi

Pith tools

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