pith. sign in

arxiv: 2605.20871 · v1 · pith:Z5U67ZBXnew · submitted 2026-05-20 · 🌌 astro-ph.GA

Robustness Analysis of USmorph: II. Optimizing Feature Extraction, Dimensionality Reduction, and Clustering for Unsupervised Galaxy Morphology Classification

Guanwen Fang , Xiaolei Yin , Yirui Zheng , Zesen Lin , Shiwei Zhu , Jie Song , Chichun Zhou , Xu Kong This is my paper

Pith reviewed 2026-05-21 03:47 UTC · model grok-4.3

classification 🌌 astro-ph.GA
keywords galaxy morphologyunsupervised classificationfeature extractiondimensionality reductionclusteringbaggingrobustnessmorphological types
0
0 comments X

The pith

Selecting AlexNet, UMAP, and bagging voting optimizes unsupervised galaxy morphology classification for better consistency and physical alignment.

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

This paper performs a robustness analysis on the unsupervised classification module of USmorph to find the best settings for processing large numbers of galaxy images. It identifies an ImageNet-pretrained AlexNet as effective for feature extraction, UMAP as suitable for reducing dimensions while keeping structure, and introduces a bagging-based voting across multiple clustering methods to make the groups more stable and pure. The optimized setup produces clusters whose distributions in parameter space match what is expected from galaxy evolution theory. Readers might value this because it offers a scalable, label-free way to categorize galaxies for big future surveys.

Core claim

The central claim is that the Bagging-based multi-cluster voting scheme significantly improves label consistency and cluster purity when combined with ImageNet-pretrained AlexNet features and UMAP dimensionality reduction, and that the resulting morphology classification results align with galaxy evolution theory by showing physically plausible distributions of different types in parameter space.

What carries the argument

Bagging-based multi-cluster voting scheme, which aggregates results from K-means, Birch, and Agglomerative clustering to enhance stability and purity of the discovered morphological groups.

If this is right

  • t-SNE plots display clear, compact cluster boundaries with strong feature separability.
  • Morphology classifications show distributions in parameter space that are consistent with galaxy evolution theory.
  • The method achieves a good balance between classification detail and the effort needed for manual checks.
  • The pipeline demonstrates sufficient robustness for application to upcoming large galaxy surveys such as CSST.

Where Pith is reading between the lines

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

  • This clustering approach might help identify unusual galaxy forms that do not fit standard categories in large datasets.
  • The stability from voting could allow the same framework to work across images from different telescopes with little adjustment.
  • If validated further, it might reduce the need for any supervised training data in initial morphology surveys.

Load-bearing premise

The features extracted using an ImageNet-pretrained AlexNet are sufficiently informative about galaxy morphological structures, and the unsupervised clusters represent genuine physical types rather than results of the specific algorithms or data choices.

What would settle it

Checking the unsupervised cluster assignments against a set of galaxies that have been classified by human experts and seeing if the purity and consistency metrics are substantially higher when using the bagging voting than when using single clustering algorithms.

Figures

Figures reproduced from arXiv: 2605.20871 by Chichun Zhou, Guanwen Fang, Jie Song, Shiwei Zhu, Xiaolei Yin, Xu Kong, Yirui Zheng, Zesen Lin.

Figure 1
Figure 1. Figure 1: Left: the redshift distribution of the selected sample; Right: the Imag distribution of the COMOS2020 in the range of 0.2 < z < 1.2 with the vertical dashed line indicating the brightness threshold Imag < 25 of the selected sample. tions, thereby improving classification accuracy in scenarios requiring rotational invariance. The APCT technique first de￾fines the initial polar axis based on the pixels with … view at source ↗
Figure 2
Figure 2. Figure 2: Flowchart of the unsupervised galaxy morphology classification pipeline of the USmorph framework, comprising three main stages: feature extraction (Panel (a)), dimensionality reduction (Panel (b)), and unsupervised clustering (Panel (c)). Net (Tan & Le 2019), ViT (Dosovitskiy et al. 2020), and ConvNext (Liu et al. 2022). Our experiments confirm that AlexNet achieves the best performance on the galaxy mor￾p… view at source ↗
Figure 3
Figure 3. Figure 3: Schematic illustration of the CAE architecture. The left part is the encoder, and the right part is the decoder. The original image is fed as input, and a denoised image is reconstructed at the output after processing through the encoder and decoder. Raw Denoised Polared Raw Denoised Polared Raw Denoised Polared 1 arcsec [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Six image sets that demonstrate image preprocessing steps. Each set contains two galaxies of the same classification category. In each set, the left, center, and right panels show the original images in the rest-frame optical band, post-CAE-based denoised images, and the images after polar coordinate expansion, respectively. The blue bar in the first panel indicates an angular scale of 1 ′′ (≈ 33 pixels). … view at source ↗
Figure 5
Figure 5. Figure 5: Comparison of total accuracy for different pretrained models and clustering methods. 0 500 1000 1500 2000 2500 3000 3500 4000 Dimension 0.0 0.2 0.4 0.6 0.8 1.0 Ratio 2800 2900 3000 0.8 1.0 [PITH_FULL_IMAGE:figures/full_fig_p009_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: The information ratio as a function of dimension with the PCA method. The ratio starts to drop significantly after 2900 di￾mensions, indicating that the main features are concentrated within the first 2900 dimensions. Therefore, we choose to reduce the di￾mensionality to 2900. purity, high-confidence clusters with substantially lower am￾biguity. These high-quality clusters serve as ready-to-use as￾sets in … view at source ↗
Figure 7
Figure 7. Figure 7: The left and middle panels display the recall and precision rates for five galaxy classes, with overall recall and precision exceeding 97%. The right panel presents the F1 scores for each galaxy class, demonstrating the UML framework’s effectiveness in distinguishing between different galaxy types. AlexNet AlexNet+Bagging AlexNet+UMAP+Bagging SPH ETD LTD IRR UNC [PITH_FULL_IMAGE:figures/full_fig_p012_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Visualization of galaxy morphological classification results in the reduced feature space using different deep learning frameworks. From left to right, we show the results of: (1) AlexNet alone, (2) AlexNet combined with Bagging, and (3) AlexNet integrated with UMAP dimensionality reduction and Bagging. Each color represents a distinct morphological class: green for SPH, purple for ETD, orange for LTD, blu… view at source ↗
Figure 9
Figure 9. Figure 9: Distribution of key morphological parameters for different galaxy classes identified by the unsupervised classification pipeline. Panels show the results of: (a) Gini coefficient G, (b) concentration index M20, (c) Sersic index ´ n, and (d) effective radius re. Each panel includes violin plots with overlaid boxplots, where the black solid line represents the median, the thick black bar indicates the interq… view at source ↗
read the original abstract

We conduct a systematic robustness analysis of the unsupervised machine learning module within the hybrid framework \texttt{USmorph}. This module automatically discovers morphological structures from large-scale galaxy images, forming the foundation of the complete classification workflow. We evaluate five pre-trained models for feature extraction and identify an ImageNet-pretrained AlexNet as the most effective for capturing discriminative morphological features. UMAP is chosen for dimensionality reduction due to its optimal balance between preserving high-dimensional structure and computational efficiency. To enhance clustering stability, we propose a Bagging-based multi-cluster voting scheme, which significantly improves label consistency and cluster purity. We compare the convergence, scalability, and quality of five clustering algorithms, finding that the Bagging voting scheme has the best performance with the combination of K-means, Birch, and Agg. A bagging clustering number of $K=16$ is used to achieve the optimal balance between classification granularity and manual validation efficiency. Our tests show that: (1) the t-distributed stochastic neighbor embedding (t-SNE) reveals clear, compact cluster boundaries in low-dimensional space with strong feature separability; (2) the morphology classification results align with galaxy evolution theory, showing physically plausible distributions of different types in parameter space. These results demonstrate the technical robustness and scientific credibility of \texttt{USmorph}, establishing it as a reliable method for automated morphological classification in future large-scale surveys such as the China Space Station Telescope (CSST) mission.

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

3 major / 2 minor

Summary. The manuscript presents a robustness analysis of the unsupervised machine learning module in the USmorph framework for automated galaxy morphology classification. It evaluates five pre-trained models for feature extraction and selects an ImageNet-pretrained AlexNet, adopts UMAP for dimensionality reduction, and introduces a Bagging-based multi-cluster voting scheme combining K-means, Birch, and Agglomerative clustering with K=16. The authors report that t-SNE visualizations demonstrate clear cluster boundaries with strong feature separability and that the resulting classifications show physically plausible distributions aligned with galaxy evolution theory, positioning the method as reliable for large-scale surveys such as CSST.

Significance. If the unsupervised clusters correspond to physically meaningful morphological types rather than algorithmic artifacts, the work could support scalable classification pipelines for upcoming surveys. However, the significance is limited by the absence of quantitative performance metrics, external validation against labeled catalogs, and tests confirming that ImageNet features prioritize morphological structures, which reduces the ability to assess scientific credibility beyond internal consistency.

major comments (3)
  1. [Abstract] Abstract: The claims that 't-SNE reveals clear, compact cluster boundaries in low-dimensional space with strong feature separability' and 'morphology classification results align with galaxy evolution theory, showing physically plausible distributions' are presented without quantitative metrics such as silhouette scores, cluster purity, adjusted Rand index, or statistical comparisons to hydrodynamic simulations or Galaxy Zoo labels.
  2. [Feature extraction] Feature extraction section: The selection of the ImageNet-pretrained AlexNet after comparing five models asserts it captures 'discriminative morphological features,' yet no analysis tests whether the 4096-dimensional vectors encode bulge-to-disk ratios, arm winding, or merger signatures versus non-morphological properties such as total flux, color, or PSF residuals.
  3. [Bagging-based multi-cluster voting scheme] Bagging scheme and clustering comparison: The choice of K=16 is justified as achieving 'the optimal balance between classification granularity and manual validation efficiency,' but this is an empirical tuning step with no reported sensitivity analysis across K values or formal derivation showing robustness of the voting scheme to this specific hyperparameter.
minor comments (2)
  1. [Abstract and Methods] The abstract and main text would benefit from explicit dataset sizes, number of galaxies analyzed, and exclusion criteria to allow reproducibility assessment.
  2. [Results] Figure captions for t-SNE visualizations should include quantitative measures of cluster separation to complement the qualitative descriptions.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for their constructive and detailed comments, which have helped us identify areas to strengthen the manuscript. We address each major comment below and outline the revisions we will make.

read point-by-point responses

  1. Referee: [Abstract] Abstract: The claims that 't-SNE reveals clear, compact cluster boundaries in low-dimensional space with strong feature separability' and 'morphology classification results align with galaxy evolution theory, showing physically plausible distributions' are presented without quantitative metrics such as silhouette scores, cluster purity, adjusted Rand index, or statistical comparisons to hydrodynamic simulations or Galaxy Zoo labels.

    Authors: We appreciate this point. The manuscript currently supports these claims through t-SNE visualizations demonstrating separability and qualitative consistency with expected galaxy evolution trends (e.g., distributions in parameter space). To enhance rigor, we will revise the abstract and add quantitative internal metrics, including silhouette scores and cluster purity values derived from the bagging scheme. However, adjusted Rand index and direct comparisons to labeled catalogs such as Galaxy Zoo or hydrodynamic simulations require supervised ground truth and are outside the scope of this unsupervised robustness analysis, which prioritizes internal stability and consistency. We will clarify this scope limitation in the revised text. revision: partial

  2. Referee: [Feature extraction] Feature extraction section: The selection of the ImageNet-pretrained AlexNet after comparing five models asserts it captures 'discriminative morphological features,' yet no analysis tests whether the 4096-dimensional vectors encode bulge-to-disk ratios, arm winding, or merger signatures versus non-morphological properties such as total flux, color, or PSF residuals.

    Authors: We thank the referee for highlighting this. Model selection was driven by comparative downstream performance: AlexNet yielded the most stable clusters and highest label consistency when paired with UMAP and the bagging voting scheme among the five tested pre-trained models. While we did not conduct an explicit interpretability study (such as feature correlations with bulge-to-disk ratios or merger indicators), the robustness results across varied galaxy samples suggest the features capture structurally relevant information. In the revision, we will add a short discussion and a supporting correlation analysis with available catalog parameters (e.g., concentration and asymmetry indices) to better substantiate that morphological content is prioritized over purely photometric or instrumental effects. revision: partial

  3. Referee: [Bagging-based multi-cluster voting scheme] Bagging scheme and clustering comparison: The choice of K=16 is justified as achieving 'the optimal balance between classification granularity and manual validation efficiency,' but this is an empirical tuning step with no reported sensitivity analysis across K values or formal derivation showing robustness of the voting scheme to this specific hyperparameter.

    Authors: We agree that a sensitivity analysis would provide stronger justification. The value K=16 was chosen empirically after testing to balance cluster granularity against the practical demands of manual validation for large surveys. In the revised manuscript, we will include a new subsection and accompanying figure that reports results for a range of K values (e.g., 8, 12, 16, 20), demonstrating how the bagging multi-cluster voting scheme maintains high label consistency and purity across these choices, thereby showing robustness to this hyperparameter. revision: yes

Circularity Check

0 steps flagged

No significant circularity; methodological choices are empirically driven and self-contained

full rationale

The paper evaluates five pre-trained models for feature extraction, selects UMAP for dimensionality reduction, and proposes a Bagging-based multi-cluster voting scheme by direct comparison of metrics including label consistency, cluster purity, convergence, and scalability. The choice of K=16 is explicitly described as an empirical trade-off for granularity versus manual validation efficiency rather than a derived quantity. Claims of alignment with galaxy evolution theory are presented as post-hoc observations of parameter-space distributions from the resulting clusters, not as predictions forced by the inputs or by self-citation chains. No equations or steps reduce by construction to fitted parameters or prior self-referential definitions; the analysis relies on independent performance benchmarks across algorithms and is therefore self-contained against external validation criteria.

Axiom & Free-Parameter Ledger

1 free parameters · 1 axioms · 0 invented entities

The central claims rest on the assumption that ImageNet features transfer to galaxy images and that unsupervised clusters map to physical morphology; K=16 is an empirically chosen free parameter.

free parameters (1)
  • bagging clustering number K = 16
    Set to 16 to balance classification granularity against manual validation effort.
axioms (1)
  • domain assumption ImageNet-pretrained convolutional networks extract features that are discriminative for galaxy morphological structures
    Invoked when declaring AlexNet the most effective model after comparing five pre-trained networks.

pith-pipeline@v0.9.0 · 5822 in / 1366 out tokens · 36031 ms · 2026-05-21T03:47:03.795962+00:00 · methodology

discussion (0)

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

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

191 extracted references · 191 canonical work pages · 17 internal anchors

  1. [2]

    , year = 1926, month = dec, volume =

    Extragalactic nebulae. , year = 1926, month = dec, volume =. doi:10.1086/143018 , adsurl =

    work page doi:10.1086/143018 1926

  2. [3]

    , keywords =

    Automating galaxy morphology classification using K nearest neighbors and non-parametric statistics. , keywords =. doi:10.1093/mnras/stae1684 , adsurl =

    work page doi:10.1093/mnras/stae1684

  3. [4]

    , keywords =

    Detecting galaxy tidal features using self-supervised representation learning. , keywords =. doi:10.1093/mnras/stae1402 , archivePrefix =. 2307.04967 , primaryClass =

    work page doi:10.1093/mnras/stae1402

  4. [5]

    Research in Astronomy and Astrophysics , abstract =

    Yu Bai and Ji-Feng Liu and Song Wang , title =. Research in Astronomy and Astrophysics , abstract =. 2018 , month =. doi:10.1088/1674-4527/18/10/118 , url =

    work page doi:10.1088/1674-4527/18/10/118 2018

  5. [7]

    Prevot, M. L. and Lequeux, J. and Maurice, E. and Prevot, L. and. The Typical Interstellar Extinction in the. 1984 , journal =

    work page 1984

  6. [8]

    Monthly Notices of the Royal Astronomical Society , volume=

    Galaxy Zoo DECaLS: Detailed visual morphology measurements from volunteers and deep learning for 314 000 galaxies , author=. Monthly Notices of the Royal Astronomical Society , volume=. 2022 , publisher=

    work page 2022

  7. [9]

    J., Jameson , A., van Straten , W., et al

    Galaxy Zoo 1: Data Release of Morphological Classifications for Nearly 900,000 Galaxies. , year = 2011, month = jan, volume =. doi:10.1111/j.1365-2966.2010.17432.x , archivePrefix =. 1007.3265 , primaryClass =

    work page doi:10.1111/j.1365-2966.2010.17432.x 2011

  8. [10]

    M., et al

    COSMOS-Web: An Overview of the JWST Cosmic Origins Survey. , keywords =. doi:10.3847/1538-4357/acc2bc , archivePrefix =. 2211.07865 , primaryClass =

    work page doi:10.3847/1538-4357/acc2bc

  9. [11]

    Early Results From GLASS-JWST. XII. The Morphology of Galaxies at the Epoch of Reionization. , keywords =. doi:10.3847/2041-8213/ac9283 , archivePrefix =. 2207.13527 , primaryClass =

    work page doi:10.3847/2041-8213/ac9283 2041

  10. [12]

    , keywords =

    A Robust Study of High-redshift Galaxies: Unsupervised Machine Learning for Characterizing Morphology with JWST up to z 8. , keywords =. doi:10.3847/1538-4357/ad17b8 , archivePrefix =. 2306.17225 , primaryClass =

    work page doi:10.3847/1538-4357/ad17b8

  11. [13]

    The Astrophysical Journal , abstract =

    Zhu, Shiwei and Fang, Guanwen and Dai, Yao and Zhou, Chichun and Zheng, Yirui and Song, Jie and Lu, Shiying and Kong, Xu , title =. The Astrophysical Journal , abstract =. 2025 , month =. doi:10.3847/1538-4357/ae1979 , url =

    work page doi:10.3847/1538-4357/ae1979 2025

  12. [14]

    , keywords =

    The JWST Hubble Sequence: The Rest-frame Optical Evolution of Galaxy Structure at 1.5 < z < 6.5. , keywords =. doi:10.3847/1538-4357/acec76 , archivePrefix =. 2210.01110 , primaryClass =

    work page doi:10.3847/1538-4357/acec76

  13. [15]

    2009 , eprint=

    LSST Science Book, Version 2.0 , author=. 2009 , eprint=

    work page 2009

  14. [16]

    2025, , 697, A1, 10.1051/0004-6361/202450810

    Euclid: I. Overview of the Euclid mission. , keywords =. doi:10.1051/0004-6361/202450810 , archivePrefix =. 2405.13491 , primaryClass =

    work page doi:10.1051/0004-6361/202450810

  15. [17]

    Science China Physics, Mechanics & Astronomy , volume=

    Introduction to the Chinese Space Station Survey Telescope (CSST) , author=. Science China Physics, Mechanics & Astronomy , volume=. 2026 , publisher=

    work page 2026

  16. [18]

    Boletin de la Asociacion Argentina de Astronomia La Plata Argentina , year = 1963, month = feb, volume =

    Influence of the atmospheric and instrumental dispersion on the brightness distribution in a galaxy. Boletin de la Asociacion Argentina de Astronomia La Plata Argentina , year = 1963, month = feb, volume =

    work page 1963

  17. [19]

    ImageNet: A large-scale hierarchical image database , year=

    Deng, Jia and Dong, Wei and Socher, Richard and Li, Li-Jia and Kai Li and Li Fei-Fei , booktitle=. ImageNet: A large-scale hierarchical image database , year=

    work page

  18. [20]

    Fully Convolutional Networks for Semantic Segmentation of Very High Resolution Remotely Sensed Images Combined With DSM , year=

    Sun, Weiwei and Wang, Ruisheng , journal=. Fully Convolutional Networks for Semantic Segmentation of Very High Resolution Remotely Sensed Images Combined With DSM , year=

    work page

  19. [21]

    The Tractor: Probabilistic astronomical source detection and measurement

    work page

  20. [22]

    , title =

    Dickinson, Hugh and Fortson, Lucy and Lintott, Chris and Scarlata, Claudia and Willett, Kyle and Bamford, Steven and Beck, Melanie and Cardamone, Carolin and Galloway, Melanie and Simmons, Brooke and Keel, William and Kruk, Sandor and Masters, Karen and Vogelsberger, Mark and Torrey, Paul and Snyder, Gregory F. , title =. The Astrophysical Journal , year ...

    work page doi:10.3847/1538-4357/aaa250

  21. [23]

    Monthly Notices of the Royal Astronomical Society , volume =

    Kolesnikov, I and Sampaio, V M and de Carvalho, R R and Conselice, C and Rembold, S B and Mendes, C L and Rosa, R R , title =. Monthly Notices of the Royal Astronomical Society , volume =. 2023 , month =. doi:10.1093/mnras/stad3934 , url =

    work page doi:10.1093/mnras/stad3934 2023

  22. [24]

    , keywords =

    The impact of morphological quenching mechanisms on star formation activity at 0.2 < z < 1.2 in the COSMOS field. , keywords =. doi:10.1051/0004-6361/202553693 , adsurl =

    work page doi:10.1051/0004-6361/202553693

  23. [25]

    and Capak, P

    Ilbert, O. and Capak, P. and Salvato, M. and Aussel, H. and McCracken, H. J. and Sanders, D. B. and Scoville, N. and Kartaltepe, J. and Arnouts, S. and Floc'h, E. Le and Mobasher, B. and Taniguchi, Y. and Lamareille, F. and Leauthaud, A. and Sasaki, S. and Thompson, D. and Zamojski, M. and Zamorani, G. and Bardelli, S. and Bolzonella, M. and Bongiorno, A....

    work page doi:10.1088/0004-637x/690/2/1236 2008

  24. [26]

    The Astrophysical Journal Supplement Series , volume=

    The COSMOS2015 Catalog: Exploring the 1< z< 6 Universe with half a million galaxies , author=. The Astrophysical Journal Supplement Series , volume=. 2016 , publisher=

    work page 2016

  25. [27]

    M., White, S

    Bruzual, G. and Charlot, S. , title = ". Monthly Notices of the Royal Astronomical Society , volume =. 2003 , month =. doi:10.1046/j.1365-8711.2003.06897.x , url =

    work page doi:10.1046/j.1365-8711.2003.06897.x 2003

  26. [28]

    Monthly Notices of the Royal Astronomical Society , volume =

    Bates, Dominic J and Tojeiro, Rita and Newman, Jeffrey A and Gonzalez-Perez, Violeta and Comparat, Johan and Schneider, Donald P and Lima, Marcos and Streblyanska, Alina , title =. Monthly Notices of the Royal Astronomical Society , volume =. 2019 , month =. doi:10.1093/mnras/stz997 , url =

    work page doi:10.1093/mnras/stz997 2019

  27. [29]

    An automatic taxonomy of galaxy morphology using unsupervised machine learning

    An automatic taxonomy of galaxy morphology using unsupervised machine learning. , keywords =. doi:10.1093/mnras/stx2351 , archivePrefix =. 1709.05834 , primaryClass =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1093/mnras/stx2351

  28. [30]

    2022 International Conference on Electrical , year = 2022, month = sep, eid =

    The Classification of Optical Galaxy Morphology Using Unsupervised Learning Techniques. 2022 International Conference on Electrical , year = 2022, month = sep, eid =. doi:10.1109/ICECET55527.2022.9872611 , archivePrefix =. 2206.06165 , primaryClass =

    work page doi:10.1109/icecet55527.2022.9872611 2022

  29. [31]

    Monthly Notices of the Royal Astronomical Society , volume =

    Cheng, Ting-Yun and Huertas-Company, Marc and Conselice, Christopher J and Aragón-Salamanca, Alfonso and Robertson, Brant E and Ramachandra, Nesar , title = ". Monthly Notices of the Royal Astronomical Society , volume =. 2021 , month =. doi:10.1093/mnras/stab734 , url =

    work page doi:10.1093/mnras/stab734 2021

  30. [32]

    Accurate photometric redshifts for the CFHT Legacy Survey calibrated using the VIMOS VLT Deep Survey

    Accurate photometric redshifts for the CFHT legacy survey calibrated using the VIMOS VLT deep survey. , keywords =. doi:10.1051/0004-6361:20065138 , archivePrefix =. astro-ph/0603217 , primaryClass =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1051/0004-6361:20065138

  31. [33]

    The Relationship Between Stellar Light Distributions of Galaxies and their Formation Histories

    The Relationship between Stellar Light Distributions of Galaxies and Their Formation Histories. , keywords =. doi:10.1086/375001 , archivePrefix =. astro-ph/0303065 , primaryClass =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1086/375001

  32. [34]

    2020 , publisher =

    Chen, Ting and Kornblith, Simon and Norouzi, Mohammad and Hinton, Geoffrey , title =. 2020 , publisher =

    work page 2020

  33. [35]

    Momentum Contrast for Unsupervised Visual Representation Learning , year=

    He, Kaiming and Fan, Haoqi and Wu, Yuxin and Xie, Saining and Girshick, Ross , booktitle=. Momentum Contrast for Unsupervised Visual Representation Learning , year=

    work page

  34. [36]

    2019 , eprint=

    Learning deep representations by mutual information estimation and maximization , author=. 2019 , eprint=

    work page 2019

  35. [37]

    Astronomy & Astrophysics , year =

    Guanwen Fang and Yao Dai and Zesen Lin and Chichun Zhou and Jie Song and Yizhou Gu and Xiaotong Guo and Anqi Mao and Xu Kong , title =. Astronomy & Astrophysics , year =. doi:10.1051/0004-6361/202451734 , note =

    work page doi:10.1051/0004-6361/202451734

  36. [38]

    Gradient Pattern Analysis Applied to Galaxy Morphology

    Gradient pattern analysis applied to galaxy morphology. , keywords =. doi:10.1093/mnrasl/sly054 , archivePrefix =. 1803.10853 , primaryClass =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1093/mnrasl/sly054

  37. [39]

    Yuan, Li and Chen, Yunpeng and Wang, Tao and Yu, Weihao and Shi, Yujun and Jiang, Zihang and Tay, Francis E. H. and Feng, Jiashi and Yan, Shuicheng , booktitle=. Tokens-to-Token ViT: Training Vision Transformers from Scratch on ImageNet , year=

    work page

  38. [40]

    ArXiv , year=

    Unsupervised Learning of Visual Features by Contrasting Cluster Assignments , author=. ArXiv , year=

    work page

  39. [41]

    ConvNeXt V2: Co-designing and Scaling ConvNets with Masked Autoencoders , year=

    Woo, Sanghyun and Debnath, Shoubhik and Hu, Ronghang and Chen, Xinlei and Liu, Zhuang and Kweon, In So and Xie, Saining , booktitle=. ConvNeXt V2: Co-designing and Scaling ConvNets with Masked Autoencoders , year=

    work page

  40. [42]

    ArXiv , year=

    An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale , author=. ArXiv , year=

    work page

  41. [44]

    , keywords =

    Sloan Digital Sky Survey: Early Data Release. , keywords =. doi:10.1086/324741 , adsurl =

    work page doi:10.1086/324741

  42. [45]

    ArXiv , year=

    UMAP: Uniform Manifold Approximation and Projection for Dimension Reduction , author=. ArXiv , year=

    work page

  43. [46]

    Law and Charles C

    David R. Law and Charles C. Steidel and Dawn K. Erb and Max Pettini and Naveen A. Reddy and Alice E. Shapley and Kurt L. Adelberger and David J. Simenc , title =. The Astrophysical Journal , abstract =. 2007 , month =. doi:10.1086/510357 , url =

    work page doi:10.1086/510357 2007

  44. [47]

    zCOSMOS - 10k-bright spectroscopic sample. The bimodality in the Galaxy Stellar Mass Function: exploring its evolution with redshift

    zCOSMOS - 10k-bright spectroscopic sample. The bimodality in the galaxy stellar mass function: exploring its evolution with redshift. , keywords =. doi:10.1051/0004-6361/200913020 , archivePrefix =. 0907.5416 , primaryClass =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1051/0004-6361/200913020

  45. [48]

    , keywords =

    COSMOS2020: A Panchromatic View of the Universe to z ∼ 10 from Two Complementary Catalogs , journal =. 2022 , month =. doi:10.3847/1538-4365/ac3078 , url =

    work page doi:10.3847/1538-4365/ac3078 2022

  46. [49]

    2019 , issue_date =

    Wu, Zifeng and Shen, Chunhua and van den Hengel, Anton , title =. 2019 , issue_date =. doi:10.1016/j.patcog.2019.01.006 , journal =

    work page doi:10.1016/j.patcog.2019.01.006 2019

  47. [50]

    2022 , eprint=

    Networks with pixels embedding: a method to improve noise resistance in images classification , author=. 2022 , eprint=

    work page 2022

  48. [51]

    and Wu, Yonghui and Prabhavalkar, Rohit and Nguyen, Patrick and Chen, Zhifeng and Kannan, Anjuli and Weiss, Ron J

    Chiu, Chung-Cheng and Sainath, Tara N. and Wu, Yonghui and Prabhavalkar, Rohit and Nguyen, Patrick and Chen, Zhifeng and Kannan, Anjuli and Weiss, Ron J. and Rao, Kanishka and Gonina, Ekaterina and Jaitly, Navdeep and Li, Bo and Chorowski, Jan and Bacchiani, Michiel , booktitle =. State-of-the-Art Speech Recognition with Sequence-to-Sequence Models , year =

    work page

  49. [52]

    1996a, A&AS, 117, 393, doi: 10.1051/aas:1996164 —

    SExtractor: Software for source extraction , DOI= "10.1051/aas:1996164", url= "https://doi.org/10.1051/aas:1996164", journal =

    work page doi:10.1051/aas:1996164

  50. [53]

    B., & Gunn, J

    Secondary standard stars for absolute spectrophotometry. , keywords =. doi:10.1086/160817 , adsurl =

    work page doi:10.1086/160817

  51. [54]

    Galactic Stellar and Substellar Initial Mass Function

    Gilles Chabrier , title =. 2003 , month =. doi:10.1086/376392 , url =

    work page internal anchor Pith review doi:10.1086/376392 2003

  52. [55]

    J., Kelly, D

    Marcia J. Rieke and Douglas M. Kelly and Karl Misselt and John Stansberry and Martha Boyer and Thomas Beatty and Eiichi Egami and Michael Florian and Thomas P. Greene and Kevin Hainline and Jarron Leisenring and Thomas Roellig and Everett Schlawin and Fengwu Sun and Lee Tinnin and Christina C. Williams and Christopher N. A. Willmer and Debra Wilson and Ch...

    work page doi:10.1088/1538-3873/acac53 2023

  53. [56]

    Space Telescopes and Instrumentation 2022: Optical, Infrared, and Millimeter Wave , year = 2022, editor =

    James Webb Space Telescope MIRI shear pupil analysis. Space Telescopes and Instrumentation 2022: Optical, Infrared, and Millimeter Wave , year = 2022, editor =. doi:10.1117/12.2632087 , adsurl =

    work page doi:10.1117/12.2632087 2022

  54. [57]

    and Drory, Niv , year =

    Fisher, David B. and Drory, Niv , year =. The Astronomical Journal , volume =. doi:10.1088/0004-6256/136/2/773 , urldate =

    work page doi:10.1088/0004-6256/136/2/773

  55. [58]

    Physical properties and environments of nearby galaxies

    Physical Properties and Environments of Nearby Galaxies. , keywords =. doi:10.1146/annurev-astro-082708-101734 , archivePrefix =. 0908.3017 , primaryClass =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1146/annurev-astro-082708-101734

  56. [59]

    Monthly Notices of the Royal Astronomical Society , volume =

    Wilde, Joshua and Serjeant, Stephen and Bromley, Jane M and Dickinson, Hugh and Koopmans, Léon V E and Metcalf, R Benton , title =. Monthly Notices of the Royal Astronomical Society , volume =. 2022 , month =. doi:10.1093/mnras/stac562 , url =

    work page doi:10.1093/mnras/stac562 2022

  57. [60]

    Monthly Notices of the Royal Astronomical Society , volume =

    Ćiprijanović, A and Kafkes, D and Downey, K and Jenkins, S and Perdue, G N and Madireddy, S and Johnston, T and Snyder, G F and Nord, B , title =. Monthly Notices of the Royal Astronomical Society , volume =. 2021 , month =. doi:10.1093/mnras/stab1677 , url =

    work page doi:10.1093/mnras/stab1677 2021

  58. [61]

    Research in Astronomy and Astrophysics , abstract =

    Shiliang Zhang and Guanwen Fang and Jie Song and Ran Li and Yizhou Gu and Zesen Lin and Chichun Zhou and Yao Dai and Xu Kong , title =. Research in Astronomy and Astrophysics , abstract =. 2024 , month =. doi:10.1088/1674-4527/ad6fe6 , url =

    work page doi:10.1088/1674-4527/ad6fe6 2024

  59. [62]

    Meddelanden fran Lunds Astronomiska Observatorium Serie II , year = 1958, month = jan, volume =

    A photographic photometry of extragalactic nebulae. Meddelanden fran Lunds Astronomiska Observatorium Serie II , year = 1958, month = jan, volume =

    work page 1958

  60. [63]

    , keywords =

    Galaxy morphology in rich clusters: implications for the formation and evolution of galaxies. , keywords =. doi:10.1086/157753 , adsurl =

    work page doi:10.1086/157753

  61. [64]

    and Mortlock, Alice and Bluck, Asa F

    Conselice, Christopher J. and Mortlock, Alice and Bluck, Asa F. L. and Grützbauch, Ruth and Duncan, Kenneth , title =. Monthly Notices of the Royal Astronomical Society , volume =. 2013 , month =. doi:10.1093/mnras/sts682 , url =

    work page doi:10.1093/mnras/sts682 2013

  62. [65]

    Kauffmann, Guinevere and White, Simon D. M. and Heckman, Timothy M. and M. The Environmental Dependence of the Relations between Stellar Mass, Structure, Star Formation and Nuclear Activity in Galaxies:. 2004 , journal =. doi:10.1111/j.1365-2966.2004.08117.x , urldate =

    work page doi:10.1111/j.1365-2966.2004.08117.x 2004

  63. [66]

    2014 , journal =

    The Connection between Galaxy Structure and Quenching Efficiency , author =. 2014 , journal =. doi:10.1093/mnras/stu331 , urldate =

    work page doi:10.1093/mnras/stu331 2014

  64. [67]

    Megan and Simmons, Brooke D

    Schawinski, Kevin and Urry, C. Megan and Simmons, Brooke D. and Fortson, Lucy and Kaviraj, Sugata and Keel, William C. and Lintott, Chris J. and Masters, Karen L. and Nichol, Robert C. and Sarzi, Marc and Skibba, Ramin and Treister, Ezequiel and Willett, Kyle W. and Wong, O. Ivy and Yi, Sukyoung K. , year =. The Green Valley Is a Red Herring:. Monthly Not...

    work page doi:10.1093/mnras/stu327

  65. [68]

    Simmons, B. D. and Melvin, Thomas and Lintott, Chris and Masters, Karen L. and Willett, Kyle W. and Keel, William C. and Smethurst, R. J. and Cheung, Edmond and Nichol, Robert C. and Schawinski, Kevin and Rutkowski, Michael and Kartaltepe, Jeyhan S. and Bell, Eric F. and Casteels, Kevin R. V. and Conselice, Christopher J. and Almaini, Omar and Ferguson, H...

    work page doi:10.1093/mnras/stu1817 2014

  66. [69]

    2019 , journal =

    Dust Properties and Star Formation of Approximately a Thousand Local Galaxies , author =. 2019 , journal =. doi:10.1051/0004-6361/201834553 , urldate =

    work page doi:10.1051/0004-6361/201834553 2019

  67. [70]

    and Glazebrook, Karl and Kacprzak, Glenn G

    Kawinwanichakij, Lalitwadee and Papovich, Casey and Quadri, Ryan F. and Glazebrook, Karl and Kacprzak, Glenn G. and Allen, Rebecca J. and Bell, Eric F. and Croton, Darren J. and Dekel, Avishai and Ferguson, Henry C. and Forrest, Ben and Grogin, Norman A. and Guo, Yicheng and Kocevski, Dale D. and Koekemoer, Anton M. and Labb. Effect of. 2017 , journal =. ...

    work page doi:10.3847/1538-4357/aa8b75 2017

  68. [71]

    , author=

    A new classification system for galaxies. , author=. The Astrophysical Journal , year=

    work page

  69. [72]

    Handbuch der Physik , year = 1959, month = jan, volume =

    Astrophysik IV: Sternsysteme / Astrophysics IV: Stellar Systems. Handbuch der Physik , year = 1959, month = jan, volume =. doi:10.1007/978-3-642-45932-0 , adsurl =

    work page doi:10.1007/978-3-642-45932-0 1959

  70. [73]

    The Morphologies of Distant Galaxies. II. Classifications from the Hubble Space Telescope Medium Deep Survey. , keywords =. doi:10.1086/192352 , adsurl =

    work page doi:10.1086/192352

  71. [74]

    Conselice and Kenneth Duncan and Ting-Yun Cheng and Alex Griffiths and Amy Whitney , title =

    Leonardo Ferreira and Christopher J. Conselice and Kenneth Duncan and Ting-Yun Cheng and Alex Griffiths and Amy Whitney , title =. The Astrophysical Journal , abstract =. 2020 , month =. doi:10.3847/1538-4357/ab8f9b , url =

    work page doi:10.3847/1538-4357/ab8f9b 2020

  72. [75]

    Rotation-Invariant Latent Semantic Representation Learning for Object Detection in VHR Optical Remote Sensing Images , year=

    Yao, Xiwen and Feng, Xiaoxu and Cheng, Gong and Han, Junwei and Guo, Lei , booktitle=. Rotation-Invariant Latent Semantic Representation Learning for Object Detection in VHR Optical Remote Sensing Images , year=

    work page

  73. [76]

    The Morphological Diversities Among Star-forming Galaxies at High Redshifts in the Great Observatories Origins Deep Survey (GOODS)

    The Morphological Diversities among Star-forming Galaxies at High Redshifts in the Great Observatories Origins Deep Survey. , keywords =. doi:10.1086/507016 , archivePrefix =. astro-ph/0606696 , primaryClass =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1086/507016

  74. [77]

    2007, ApJS, 172, 1, doi: 10.1086/516585

    Scoville, N. and Aussel, H. and Brusa, M. and Capak, P. and Carollo, C. M. and Elvis, M. and Giavalisco, M. and Guzzo, L. and Hasinger, G. and Impey, C. and Kneib, J.‐P. and LeFevre, O. and Lilly, S. J. and Mobasher, B. and Renzini, A. and Rich, R. M. and Sanders, D. B. and Schinnerer, E. and Schminovich, D. and Shopbell, P. and Taniguchi, Y. and Tyson, N...

    work page doi:10.1086/516585

  75. [78]

    Schmidhuber, ”Deep Learning in Neural Networks: An Overview,” Neural Networks 61, 85-117 (2015).doi:10.1016/j.neunet.2014.09.003

    Schmidhuber, Jürgen , year=. Deep learning in neural networks: An overview , volume=. doi:10.1016/j.neunet.2014.09.003 , journal=

    work page doi:10.1016/j.neunet.2014.09.003 2014

  76. [79]

    and Dambre, Joni , year=

    Dieleman, Sander and Willett, Kyle W. and Dambre, Joni , year=. Rotation-invariant convolutional neural networks for galaxy morphology prediction , volume=. Monthly Notices of the Royal Astronomical Society , publisher=. doi:10.1093/mnras/stv632 , number=

    work page doi:10.1093/mnras/stv632

  77. [80]

    Galaxy morphology classification with deep convolutional neural networks , volume =

    Zhu, Xiao-Pan and Dai, Jia-Ming and Bian, Chun-Jiang and Chen, Yu and Chen, Shi and Hu, Chen , year =. Galaxy morphology classification with deep convolutional neural networks , volume =. Astrophysics and Space Science , doi =

    work page

  78. [81]

    Galaxy Morphology Classification Based on VGG19 Deep Convolutional Neural Network , year=

    Fernando, Thrinith and Rathnayake, Samadhi and Dissanayaka, Kapila , booktitle=. Galaxy Morphology Classification Based on VGG19 Deep Convolutional Neural Network , year=

    work page

  79. [82]

    2021 , eprint=

    An Unsupervised Deep-Learning Method for Fingerprint Classification: the CCAE Network and the Hybrid Clustering Strategy , author=. 2021 , eprint=

    work page 2021

  80. [83]

    Monthly Notices of the Royal Astronomical Society , volume =

    Ye, Renhao and Shen, Shiyin and de Souza, Rafael S and Xu, Quanfeng and Chen, Mi and Chen, Zhu and Ishida, Emille E O and Krone-Martins, Alberto and Durgesh, Rupesh , title =. Monthly Notices of the Royal Astronomical Society , volume =. 2025 , month =. doi:10.1093/mnras/staf025 , url =

    work page doi:10.1093/mnras/staf025 2025

Showing first 80 references.