Pith. sign in

REVIEW 4 major objections 3 minor 3 cited by

The ALMA-QUARKS Survey: III. Clump-to-core fragmentation and search for high-mass starless cores

T0 review · 4 major / 3 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read Two starless cores above 16 solar masses are scarce, favoring competitive accretion in IR-bright protoclusters.

desk verdict The abstract advertises a potentially valuable large-sample ALMA survey, but the attached full text is an unrelated computer-vision paper, leaving the claims unauditable. read the letter →

arxiv 2508.03229 v1 pith:WIWEY5SO submitted 2025-08-05 astro-ph.GA

Dongting Yang , Hong-Li Liu , Tie Liu , Xunchuan Liu , Fengwei Xu , Sheng-Li Qin , Anandmayee Tej , Guido Garay
show 38 more authors
Lei Zhu Xiaofeng Mai Wenyu Jiao Siju Zhang Sami Dib Amelia M. Stutz Aina Palau Patricio Sanhueza Annie Zavagno A.Y. Yang Xindi Tang Mengyao Tang Yichen Zhang Pablo Garcia Tianwei Zhang Anindya Saha Shanghuo Li Paul F. Goldsmith Leonardo Bronfman Chang Won Lee Kotomi Taniguchi Swagat Ranjan Das Prasanta Gorai Ariful Hoque Li Chen Zhiping Kou Jianjun Zhou Yankun Zhang L. Viktor Toth Tapas Baug Xianjin Shen Chuanshou Li Jiahang Zou Ankan Das Hafiz Nazeer L. K. Dewangan Jihye Hwang James O. Chibueze
This is my paper
classification astro-ph.GA
keywords massivestarformationstarlesscoresALMAclumpfragmentationJeanslengthcompetitiveaccretionprotoclusters1.3mmcontinuum
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 the seeds of high-mass stars exist as massive, quiescent cores before star formation begins. Using ALMA 1.3 mm observations of 139 infrared-bright massive protocluster clumps, it identifies 1,562 compact cores and classifies them by signs of star formation. The authors find that while most cores show warm or evolved signatures, only two starless cores exceed 16 solar masses. They interpret this scarcity as evidence against the turbulent core accretion picture, in which high-mass stars form from pre-existing massive cores, and in favor of competitive accretion, where mass is gathered during the formation process.

What carries the argument

The analysis uses getsf to extract compact cores from 1.3 mm continuum maps built from combined ALMA 12-m and ACA 7-m data, reaching about 0.02 pc resolution and about 0.3 $M_{\odot}$ sensitivity at 30 K. Cores are classified as starless, warm, or evolved based on associated outflow and ionized gas signatures. The comparison of observed core separations to the thermal Jeans length partitions the sample and carries the fragmentation argument.

What would settle it

Point deeper ALMA or JWST observations at the two >16 $M_{\odot}$ starless candidates to search for weak outflows, compact HII regions, or hot molecular cores; detection of any of these would show the cores are not truly starless and would remove the two candidates, while a sensitive non-detection across all starless cores would strengthen the paper's conclusion.

Watch

Extended reading notes

Core claim

The paper's central claim is that within IR-bright protocluster clumps, high-mass starless cores are rare, so high-mass star formation is unlikely to proceed through the monolithic collapse of a single massive core. Instead, the observed dearth of >16 $M_{\odot}$ starless cores favors competitive accretion-type models in which protostars gain their final mass by accreting from the surrounding clump environment. This conclusion rests on a census of 1,562 cores, of which only 127 are starless and only two exceed 16 $M_{\odot}$, combined with the finding that typical core separations are only about one-fifth of the thermal Jeans length, indicating ongoing fragmentation.

Load-bearing premise

The classification of a core as starless depends on the absence of outflow and ionized gas signatures; if embedded protostars are hidden below the sensitivity or angular resolution, some apparently starless cores would actually be evolved, and the census that the model comparison depends on would change.

Editorial extensions

If this is right

  • Observed core separations are significantly smaller than the thermal Jeans length, with the ratio peaking at about 0.2, indicating that thermal Jeans fragmentation has occurred within these clumps.
  • Among 1,562 cores, only 127 are classified as starless, while 971 are warm and 464 are evolved, showing that the vast majority of cores already host or show signs of star formation.
  • Only two starless cores have masses exceeding 16 $M_{\odot}$, so high-mass starless cores are rare in IR-bright protocluster clumps.
  • This scarcity supports competitive accretion-type models over turbulent core accretion-type models for high-mass star formation in IR-bright environments.
  • The combined ALMA 12-m and ACA data provide the angular resolution and sensitivity needed to build a clump-to-core fragmentation census at about 0.02 pc scales.

Reading between the lines

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

  • The low observed-to-Jeans length ratio could also be produced by hierarchical fragmentation, where cores form inside larger fragments; mapping the spatial distribution of the 127 starless cores across individual clumps would test this alternative.
  • If the scarcity is real, then searches for high-mass starless cores in quieter, less IR-bright clumps may find a higher abundance, which would suggest the competitive-accretion conclusion is specific to IR-bright protocluster environments.
  • The mass completeness limit near 0.3 $M_{\odot}$ implies that many lower-mass cores are missed; correcting for incompleteness could raise the total core count but is unlikely to change the dearth of >16 $M_{\odot}$ starless candidates unless sensitivity is strongly mass-dependent.
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

4 major / 3 minor

Summary. The manuscript as received presents an abstract for the ALMA-QUARKS Survey III paper, claiming to identify 1562 compact cores in 139 IR-bright massive protoclusters, to measure a clump-to-core fragmentation ratio (lambda_obs/lambda_J peaking near 0.2), to classify cores as starless/warm/evolved (127/971/464), and to find two starless cores above 16 Msun. On this basis it argues that competitive accretion-type models may be more applicable than turbulent core accretion-type models in IR-bright protocluster clumps. However, the submitted full text is an unrelated computer-vision manuscript, 'Trace3D: Consistent Segmentation Lifting via Gaussian Instance Tracing' (arXiv:2508.03227v1). None of the methods, data reduction, catalog, figures, tables, or uncertainty analysis for the ALMA study are present. The central quantitative claims and the model-discrimination conclusion are therefore unauditable as submitted.

Significance. If the abstract's results could be verified, the paper would provide a large, uniform ALMA survey of fragmentation in IR-bright protoclusters, with a specific observational constraint on the abundance of high-mass starless cores and a direct bearing on the competitive accretion versus turbulent core accretion debate. The classification scheme and the push to identify high-mass starless cores are of clear interest to the massive star formation community. However, because the body of the manuscript is absent, none of these contributions can currently be assessed; the paper cannot be assigned significance without the underlying methods and data.

major comments (4)
  1. [Full Text (manuscript body)] The submitted full text is not the ALMA-QUARKS III paper; it is the Trace3D computer-vision paper (arXiv:2508.03227v1) about Gaussian Splatting segmentation. The abstract describes an ALMA study, but the body contains no methods, observations, extraction parameters, tables, or figures relevant to that study. Every quantitative result in the abstract (1562 cores from getsf, lambda_obs/lambda_J peaking at ~0.2, 127/971/464 starless/warm/evolved split, two starless cores exceeding 16 Msun) is therefore unsupported. This is a load-bearing evidentiary gap that prevents any verification of the central claim. The authors must resubmit with the correct full text before the paper can be reviewed.
  2. [Abstract (fragmentation ratio)] The abstract states that lambda_obs/lambda_J peaks at ~0.2 and that 'thermal Jeans fragmentation has taken place,' but no uncertainties are given for this ratio, no completeness limits are described, and the computation of lambda_J (e.g., assumed temperature, density, and whether the Jeans length is computed from clump-averaged or core-scale quantities) is not specified. Without these details, the peak value cannot be interpreted as evidence of Jeans fragmentation or of hierarchical fragmentation; both interpretations are claimed in the same paragraph.
  3. [Abstract (starless classification)] The core classification into 127 starless, 971 warm, and 464 evolved cores relies on 'associated signatures of star formation (e.g., outflows and ionized gas).' The abstract does not state the sensitivity or resolution limits of the outflow/ionized-gas tracers. If embedded protostars are undetected in some fraction of the apparently starless cores, then the number of high-mass starless cores could be overestimated or underestimated, depending on the direction of contamination. Because the final model comparison depends specifically on the count of starless cores above 16 Msun (two), this classification is load-bearing and must be justified with detection thresholds.
  4. [Abstract (model discrimination)] The concluding inference — that the scarcity of high-mass starless core candidates 'suggests that competitive accretion-type models could be more applicable than turbulent core accretion-type models' — is a model-discrimination claim that requires a quantitative expectation for how many massive starless cores each model predicts in this sample, given the selection of IR-bright clumps, the mass sensitivity, and the physical assumptions. None of these elements are present in the abstract, and the body that would contain them is missing. As it stands, the inference is an interpretive leap from a single number (two) without statistical context.
minor comments (3)
  1. [Abstract (units)] The abstract gives the continuum sensitivity as '~0.6 mJy beam^-1 (~0.3 Msun at 30 K)'; the implied distance should be stated explicitly in the same sentence, since the mass conversion depends on distance and dust opacity assumptions.
  2. [Full Text (typographical)] The wrong manuscript contains several typographical errors (e.g., 'Gaussains' in Section 5, 'wth' in Figure 7 caption, 'Tab. S.7 Tab. S.8' missing comma). I do not list them in detail because this text is not the submitted ALMA paper; they should be ignored after the correct full text is supplied.
  3. [General] The abstract should define 'IR-bright' quantitatively (e.g., selection criteria on mid-IR/flux thresholds) so that the sample is reproducible.

Circularity Check

0 steps flagged · score 0.0 of 10

No circular derivation found; the astronomy abstract is an empirical inference, and the supplied full text is an unrelated paper that cannot be audited.

full rationale

The only astronomical content available is the abstract of arXiv:2508.03229. Its derivation chain is: getsf extraction of 1562 cores from 1.3 mm ALMA+ACA continuum; measurement of core separations lambda_obs relative to the Jeans length lambda_J; classification into 127 starless, 971 warm, and 464 evolved cores based on outflow and ionized-gas signatures; identification of two starless cores above 16 Msun; and an interpretive comparison between competitive accretion and turbulent core accretion models. None of these steps is defined in terms of its conclusion, and no fitted parameter is renamed as a prediction. The starless classification is an input classification, not a consequence of the model preference, and the scarcity of high-mass starless cores is an empirical result that does not by construction force the competitive-accretion conclusion. The notable anomaly is that the supplied full text is Trace3D, a computer-vision paper on Gaussian Splatting, rather than the ALMA-QUARKS III body; this makes the astronomy claims unauditable, but an evidentiary gap is not circularity and does not satisfy the requirement to exhibit a specific reduction such as Eq. X = Eq. Y or a fitted parameter presented as a prediction. Accordingly the circularity score is 0.

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

Because the received full text is an unrelated manuscript, the ledger entries are reconstructed from the abstract alone and may not match the actual paper's modeling choices.

free parameters (3)
  • Assumed dust temperature for mass estimates = 30 K
    Used to convert 1.3 mm flux to core mass and to quote the 0.3 solar mass sensitivity; core masses and the high-mass starless candidates depend on this single assumed temperature.
  • Reference distance for physical scales = 3.7 kpc (average)
    The abstract converts angular resolution to 0.02 pc using a representative distance; individual clump distances vary and affect both Jeans length and core mass.
  • Core detection threshold in getsf = Not stated
    The number of cores and their separations depend on the source extraction parameters; the abstract does not report the detection threshold or completeness limit.
assumptions (3)
  • domain assumption The thermal Jeans length is the appropriate fragmentation scale for these clumps.
    The interpretation of lambda_obs/lambda_J requires that thermal pressure and gravity set the reference scale; turbulence and magnetic fields are not included in that reference.
  • domain assumption Absence of outflow and ionized gas signatures means a core is truly starless.
    Cores are classified as starless if no star formation signatures are seen; this assumes such tracers would be detected and are reliable, which directly affects the count of high-mass starless cores.
  • domain assumption The sample is complete for high-mass starless cores at the quoted sensitivity and resolution.
    The scarcity argument depends on not missing massive starless cores due to sensitivity, resolution, or source extraction; the abstract does not provide completeness limits.

how reviews work

0 comments
Cite this review

Pith. "Pith review of The ALMA-QUARKS Survey: III. Clump-to-core fragmentation and search for high-mass starless cores." pith.science (2026). https://pith.science/paper/WIWEY5SO

@misc{pith2026250803229,
  author       = {Pith},
  title        = {Pith review of: The ALMA-QUARKS Survey: III. Clump-to-core fragmentation and search for high-mass starless cores},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/WIWEY5SO}},
  note         = {Machine review of arXiv:2508.03229}
}
abstract

The Querying Underlying mechanisms of massive star formation with ALMA-Resolved gas Kinematics and Structures (QUARKS) survey observed 139 infrared-bright (IR-bright) massive protoclusters at 1.3 mm wavelength with ALMA. This study investigates clump-to-core fragmentation and searches for candidate high-mass starless cores within IR-bright clumps using combined ALMA 12-m (C-2) and Atacama Compact Array (ACA) 7-m data, providing $\sim$ 1 arcsec ($\sim\rm0.02~pc$ at 3.7 kpc) resolution and $\sim\rm0.6\,mJy\,beam^{-1}$ continuum sensitivity ($\sim 0.3~M_{\odot}$ at 30 K). We identified 1562 compact cores from 1.3 mm continuum emission using getsf. Observed linear core separations ($\lambda_{\rm obs}$) are significantly less than the thermal Jeans length ($\lambda_{\rm J}$), with the $\lambda_{\rm obs}/\lambda_{\rm J}$ ratios peaking at $\sim0.2$. This indicates that thermal Jeans fragmentation has taken place within the IR-bright protocluster clumps studied here. The observed low ratio of $\lambda_{\rm obs}/\lambda_{\rm J}\ll 1$ could be the result of evolving core separation or hierarchical fragmentation. Based on associated signatures of star formation (e.g., outflows and ionized gas), we classified cores into three categories: 127 starless, 971 warm, and 464 evolved cores. Two starless cores have mass exceeding 16$\,M_{\odot}$, and represent high-mass candidates. The scarcity of such candidates suggests that competitive accretion-type models could be more applicable than turbulent core accretion-type models in high-mass star formation within these IR-bright protocluster clumps.

Discussion (0). Sign in to comment.

Forward citations

Cited by 3 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Challenges in probing turbulent and magnetic support in cores: the W43-MM1 protocluster case study

    astro-ph.GA 2026-07 conditional novelty 6.0 of 10

    Simplified virial analyses of W43-MM1 cores overestimate non-thermal support because linewidths include organized motions of 1–3 km/s and surface terms are omitted, producing unexpectedly high stability fractions.

  2. How Should We Understand the Core Mass Function? A memo of the CMF2IMF conference at ESO Garching

    astro-ph.GA 2026-07 conditional novelty 5.5 of 10

    High-mass CMF slopes depend strongly on minimum fitting mass; early-stage cores (ASHES) appear steeper, consistent with an evolving high-mass end.

  3. How Should We Understand the Core Mass Function? A memo of the CMF2IMF conference at ESO Garching

    astro-ph.GA 2026-07 conditional novelty 4.0 of 10

    The high-mass slope of the core mass function depends strongly on the minimum fitting mass: completeness-based fits look top-heavy, while KS-selected tail fits move toward Salpeter, and the early-stage ASHES sample ap...

Reference graph

Works this paper leans on

72 extracted references · 60 canonical work pages · cited by 2 Pith papers

  1. [1]

    Barron, Ben Mildenhall, Matthew Tancik, Peter Hedman, Ricardo Martin-Brualla, and Pratul P

    Jonathan T. Barron, Ben Mildenhall, Matthew Tancik, Peter Hedman, Ricardo Martin-Brualla, and Pratul P. Srinivasan. Mip-nerf: A multiscale representation for anti-aliasing neu- ral radiance fields. In International Conference on Computer Vision (ICCV), 2021

    work page 2021

  2. [2]

    Barron, Ben Mildenhall, Dor Verbin, Pratul P

    Jonathan T. Barron, Ben Mildenhall, Dor Verbin, Pratul P. Srinivasan, and Peter Hedman. Zip-nerf: Anti-aliased grid- based neural radiance fields. In International Conference on Computer Vision (ICCV), 2023

    work page 2023

  3. [3]

    Contrastive lift: 3d object instance segmentation by slow-fast contrastive fusion

    Yash Bhalgat, Iro Laina, João F Henriques, Andrew Zisser- man, and Andrea Vedaldi. Contrastive lift: 3d object instance segmentation by slow-fast contrastive fusion. InAdvances in Neural Information Processing Systems (NeurIPS), 2023

    work page 2023

  4. [4]

    Dm-nerf: 3d scene geometry decomposition and manipulation from 2d images

    Wang Bing, Lu Chen, and Bo Yang. Dm-nerf: 3d scene geometry decomposition and manipulation from 2d images. In International Conference on Learning Representations (ICLR), 2023

    work page 2023

  5. [5]

    Seg- ment anything in 3d with radiance fields

    Jiazhong Cen, Jiemin Fang, Zanwei Zhou, Chen Yang, Lingxi Xie, Xiaopeng Zhang, Wei Shen, and Qi Tian. Seg- ment anything in 3d with radiance fields. arXiv preprint arXiv:2304.12308, 2023

    arXiv 2023

  6. [6]

    Segment anything in 3d with nerfs

    Jiazhong Cen, Zanwei Zhou, Jiemin Fang, Chen Yang, Wei Shen, Lingxi Xie, Xiaopeng Zhang, and Qi Tian. Segment anything in 3d with nerfs. InAdvances in Neural Information Processing Systems (NeurIPS), 2023

    work page 2023

  7. [7]

    Lifting by Gaussians: A Simple, Fast and Flexible Method for 3D Instance Segmentation

    Rohan Chacko, Nicolai Haeni, Eldar Khaliullin, Lin Sun, and Douglas Lee. Lifting by gaussians: A simple, fast and flexible method for 3d instance segmentation. arXiv preprint arXiv:2502.00173, 2025

    work page Pith review arXiv 2025

  8. [8]

    A simple framework for contrastive learn- ing of visual representations

    Ting Chen, Simon Kornblith, Mohammad Norouzi, and Ge- offrey Hinton. A simple framework for contrastive learn- ing of visual representations. In International Conference on Machine Learning (ICML), 2020

    work page 2020

Show all 72 references
  1. [10]

    Gaussianeditor: Swift and controllable 3d editing with gaussian splatting

    Yiwen Chen, Zilong Chen, Chi Zhang, Feng Wang, Xiaofeng Yang, Yikai Wang, Zhongang Cai, Lei Yang, Huaping Liu, and Guosheng Lin. Gaussianeditor: Swift and controllable 3d editing with gaussian splatting. In Conference on Com- puter Vision and Pattern Recognition (CVPR), 2024

    2024

  2. [11]

    Single-view 3d scene reconstruc- tion with high-fidelity shape and texture

    Yixin Chen, Junfeng Ni, Nan Jiang, Yaowei Zhang, Yixin Zhu, and Siyuan Huang. Single-view 3d scene reconstruc- tion with high-fidelity shape and texture. In International Conference on 3D Vision (3DV), 2024

    2024

  3. [12]

    Schwing, and Alexander Kir- illov

    Bowen Cheng, Alexander G. Schwing, and Alexander Kir- illov. Per-pixel classification is not all you need for semantic segmentation. In Advances in Neural Information Process- ing Systems (NeurIPS), 2021

    2021

  4. [13]

    Learning segmented 3d gaussians via efficient feature unprojection for zero-shot neural scene segmenta- tion

    Bin Dou, Tianyu Zhang, Zhaohui Wang, Yongjia Ma, and Zejian Yuan. Learning segmented 3d gaussians via efficient feature unprojection for zero-shot neural scene segmenta- tion. arXiv preprint arXiv:2401.05925, 2024

    2024 arXiv

  5. [14]

    OpenNeRF: Open Set 3D Neural Scene Segmentation with Pixel-Wise Features and Rendered Novel Views

    Francis Engelmann, Fabian Manhardt, Michael Niemeyer, Keisuke Tateno, Marc Pollefeys, and Federico Tombari. OpenNeRF: Open Set 3D Neural Scene Segmentation with Pixel-Wise Features and Rendered Novel Views. In Inter- national Conference on Learning Representations (ICLR) , 2024

    2024

  6. [15]

    Panoptic nerf: 3d-to-2d label transfer for panoptic urban scene segmentation

    Xiao Fu, Shangzhan Zhang, Tianrun Chen, Yichong Lu, Lanyun Zhu, Xiaowei Zhou, Andreas Geiger, and Yiyi Liao. Panoptic nerf: 3d-to-2d label transfer for panoptic urban scene segmentation. In International Conference on 3D Vi- sion (3DV), 2022

    2022

  7. [16]

    Narayanan

    Rahul Goel, Dhawal Sirikonda, Saurabh Saini, and P.J. Narayanan. Interactive Segmentation of Radiance Fields. In Conference on Computer Vision and Pattern Recognition (CVPR), 2023

    2023

  8. [17]

    Egolifter: Open-world 3d seg- mentation for egocentric perception

    Qiao Gu, Zhaoyang Lv, Duncan Frost, Simon Green, Julian Straub, and Chris Sweeney. Egolifter: Open-world 3d seg- mentation for egocentric perception. In European Confer- ence on Computer Vision (ECCV), 2025

    2025

  9. [18]

    Sam-guided graph cut for 3d instance segmentation

    Haoyu Guo, He Zhu, Sida Peng, Yuang Wang, Yujun Shen, Ruizhen Hu, and Xiaowei Zhou. Sam-guided graph cut for 3d instance segmentation. In ECCV, 2024

    2024

  10. [19]

    Semantic gaussians: Open-vocabulary scene understanding with 3d gaussian splatting.https://arxiv.org/abs/2403.15624, 2024

    Jun Guo, Xiaojian Ma, Yue Fan, Huaping Liu, and Qing Li. Semantic gaussians: Open-vocabulary scene understanding with 3d gaussian splatting.https://arxiv.org/abs/2403.15624, 2024

    2024 arXiv

  11. [20]

    Dimension- ality reduction by learning an invariant mapping

    Raia Hadsell, Sumit Chopra, and Yann LeCun. Dimension- ality reduction by learning an invariant mapping. In Confer- ence on Computer Vision and Pattern Recognition (CVPR) , 2006

    2006

  12. [21]

    Sagd: Boundary- enhanced segment anything in 3d gaussian via gaussian de- composition

    Xu Hu, Yuxi Wang, Lue Fan, Junsong Fan, Junran Peng, Zhen Lei, Qing Li, and Zhaoxiang Zhang. Sagd: Boundary- enhanced segment anything in 3d gaussian via gaussian de- composition. arXiv preprint arXiv:2401.17857, 2024

    2024 arXiv

  13. [22]

    2d gaussian splatting for geometrically ac- curate radiance fields

    Binbin Huang, Zehao Yu, Anpei Chen, Andreas Geiger, and Shenghua Gao. 2d gaussian splatting for geometrically ac- curate radiance fields. In ACM SIGGRAPH 2024 conference papers, 2024

    2024

  14. [23]

    Point cloud labeling using 3d convolutional neural network

    Jing Huang and Suya You. Point cloud labeling using 3d convolutional neural network. In International Conference on Pattern Recognition (ICPR), 2016

    2016

  15. [24]

    An embodied generalist agent in 3d world

    Jiangyong Huang, Silong Yong, Xiaojian Ma, Xiongkun Linghu, Puhao Li, Yan Wang, Qing Li, Song-Chun Zhu, Baoxiong Jia, and Siyuan Huang. An embodied generalist agent in 3d world. In International Conference on Machine Learning (ICML), 2024

    2024

  16. [25]

    Gaus- siancut: Interactive segmentation via graph cut for 3d gaus- sian splatting

    Umangi Jain, Ashkan Mirzaei, and Igor Gilitschenski. Gaus- siancut: Interactive segmentation via graph cut for 3d gaus- sian splatting. In Advances in Neural Information Processing Systems (NeurIPS), 2024

    2024

  17. [26]

    Sceneverse: Scaling 3d vision-language learning for grounded scene understanding

    Baoxiong Jia, Yixin Chen, Huangyue Yu, Yan Wang, Xuesong Niu, Tengyu Liu, Qing Li, and Siyuan Huang. Sceneverse: Scaling 3d vision-language learning for grounded scene understanding. In European Conference on Computer Vision (ECCV), 2024

    2024

  18. [27]

    3d gaussian splatting for real-time 9 radiance field rendering

    Bernhard Kerbl, Georgios Kopanas, Thomas Leimkühler, and George Drettakis. 3d gaussian splatting for real-time 9 radiance field rendering. ACM Transactions on Graphics (TOG), 2023

    2023

  19. [28]

    Lerf: Language embedded radiance fields

    Justin* Kerr, Chung Min* Kim, Ken Goldberg, Angjoo Kanazawa, and Matthew Tancik. Lerf: Language embedded radiance fields. In International Conference on Computer Vision (ICCV), 2023

    2023

  20. [29]

    Garfield: Group anything with radiance fields

    Chung Min Kim, Mingxuan Wu, Justin Kerr, Ken Goldberg, Matthew Tancik, and Angjoo Kanazawa. Garfield: Group anything with radiance fields. In Conference on Computer Vision and Pattern Recognition (CVPR), 2024

    2024

  21. [30]

    Segment anything

    Alexander Kirillov, Eric Mintun, Nikhila Ravi, Hanzi Mao, Chloe Rolland, Laura Gustafson, Tete Xiao, Spencer White- head, Alexander C Berg, Wan-Yen Lo, et al. Segment anything. In International Conference on Computer Vision (ICCV), 2023

    2023

  22. [31]

    Decomposing nerf for editing via feature field distilla- tion

    Sosuke Kobayashi, Eiichi Matsumoto, and Vincent Sitz- mann. Decomposing nerf for editing via feature field distilla- tion. In Advances in Neural Information Processing Systems (NeurIPS), 2022

    2022

  23. [32]

    Panoptic neural fields: A semantic object-aware neural scene representation

    Abhijit Kundu, Kyle Genova, Xiaoqi Yin, Alireza Fathi, Car- oline Pantofaru, Leonidas J Guibas, Andrea Tagliasacchi, Frank Dellaert, and Thomas Funkhouser. Panoptic neural fields: A semantic object-aware neural scene representation. In Conference on Computer Vision and Pattern...

    2022

  24. [33]

    Rethinking open-vocabulary segmen- tation of radiance fields in 3d space

    Hyunjee Lee, Youngsik Yun, Jeongmin Bae, Seoha Kim, and Youngjung Uh. Rethinking open-vocabulary segmen- tation of radiance fields in 3d space. arXiv preprint arXiv:2408.07416, 2024

    2024 arXiv

  25. [34]

    Dl3dv-10k: A large-scale scene dataset for deep learning-based 3d vision

    Lu Ling, Yichen Sheng, Zhi Tu, Wentian Zhao, Cheng Xin, Kun Wan, Lantao Yu, Qianyu Guo, Zixun Yu, Yawen Lu, et al. Dl3dv-10k: A large-scale scene dataset for deep learning-based 3d vision. In Conference on Computer Vision and Pattern Recognition (CVPR), 2024

    2024

  26. [35]

    Building interactable replicas of complex articulated objects via gaussian splatting

    Yu Liu, Baoxiong Jia, Ruijie Lu, Junfeng Ni, Song-Chun Zhu, and Siyuan Huang. Building interactable replicas of complex articulated objects via gaussian splatting. In In- ternational Conference on Learning Representations (ICLR), 2025

    2025

  27. [36]

    Manigaussian: Dynamic gaus- sian splatting for multi-task robotic manipulation

    Guanxing Lu, Shiyi Zhang, Ziwei Wang, Changliu Liu, Ji- wen Lu, and Yansong Tang. Manigaussian: Dynamic gaus- sian splatting for multi-task robotic manipulation. In Euro- pean Conference on Computer Vision (ECCV), 2024

    2024

  28. [37]

    Taco: Taming diffusion for in-the-wild video amodal completion

    Ruijie Lu, Yixin Chen, Yu Liu, Jiaxiang Tang, Junfeng Ni, Diwen Wan, Gang Zeng, and Siyuan Huang. Taco: Taming diffusion for in-the-wild video amodal completion. In Inter- national Conference on Computer Vision (ICCV), 2025

    2025

  29. [38]

    Movis: En- hancing multi-object novel view synthesis for indoor scenes

    Ruijie Lu, Yixin Chen, Junfeng Ni, Baoxiong Jia, Yu Liu, Diwen Wan, Gang Zeng, and Siyuan Huang. Movis: En- hancing multi-object novel view synthesis for indoor scenes. In Conference on Computer Vision and Pattern Recognition (CVPR), 2025

    2025

  30. [39]

    Dreamart: Generating interactable articulated ob- jects from a single image

    Ruijie Lu, Yu Liu, Jiaxiang Tang, Junfeng Ni, Yuxiang Wang, Diwen Wan, Gang Zeng, Yixin Chen, and Siyuan Huang. Dreamart: Generating interactable articulated ob- jects from a single image. arXiv preprint arXiv:2507.05763, 2025

    2025 arXiv

  31. [40]

    Gaga: Group any gaussians via 3d-aware memory bank

    Weijie Lyu, Xueting Li, Abhijit Kundu, Yi-Hsuan Tsai, and Ming-Hsuan Yang. Gaga: Group any gaussians via 3d-aware memory bank. arXiv preprint arXiv:2404.07977, 2024

    2024 arXiv

  32. [41]

    Total-decom: Decomposed 3d scene recon- struction with minimal interaction

    Xiaoyang Lyu, Chirui Chang, Peng Dai, Yang-Tian Sun, and Xiaojuan Qi. Total-decom: Decomposed 3d scene recon- struction with minimal interaction. In Conference on Com- puter Vision and Pattern Recognition (CVPR), 2024

    2024

  33. [42]

    Nerf: Representing scenes as neural radiance fields for view syn- thesis

    Ben Mildenhall, Pratul P Srinivasan, Matthew Tancik, Jonathan T Barron, Ravi Ramamoorthi, and Ren Ng. Nerf: Representing scenes as neural radiance fields for view syn- thesis. Communications of the ACM, 2021

    2021

  34. [43]

    Spin-nerf: Multiview segmentation and perceptual inpainting with neural radiance fields

    Ashkan Mirzaei, Tristan Aumentado-Armstrong, Konstanti- nos G Derpanis, Jonathan Kelly, Marcus A Brubaker, Igor Gilitschenski, and Alex Levinshtein. Spin-nerf: Multiview segmentation and perceptual inpainting with neural radiance fields. In Conference on Computer Vision and Pa...

    2023

  35. [44]

    Phyrecon: Physically plausible neural scene recon- struction

    Junfeng Ni, Yixin Chen, Bohan Jing, Nan Jiang, Bin Wang, Bo Dai, Puhao Li, Yixin Zhu, Song-Chun Zhu, and Siyuan Huang. Phyrecon: Physically plausible neural scene recon- struction. In Advances in Neural Information Processing Systems (NeurIPS), 2024

    2024

  36. [45]

    Decompositional neu- ral scene reconstruction with generative diffusion prior

    Junfeng Ni, Yu Liu, Ruijie Lu, Zirui Zhou, Song-Chun Zhu, Yixin Chen, and Siyuan Huang. Decompositional neu- ral scene reconstruction with generative diffusion prior. In Conference on Computer Vision and Pattern Recognition (CVPR), 2025

    2025

  37. [46]

    Pointnet: Deep learning on point sets for 3d classification and segmentation

    Charles R Qi, Hao Su, Kaichun Mo, and Leonidas J Guibas. Pointnet: Deep learning on point sets for 3d classification and segmentation. In Conference on Computer Vision and Pattern Recognition (CVPR), 2017

    2017

  38. [47]

    Langsplat: 3d language gaussian splatting

    Minghan Qin, Wanhua Li, Jiawei Zhou, Haoqian Wang, and Hanspeter Pfister. Langsplat: 3d language gaussian splatting. In Conference on Computer Vision and Pattern Recognition (CVPR), 2024

    2024

  39. [48]

    Sam 2: Segment anything in images and videos

    Nikhila Ravi, Valentin Gabeur, Yuan-Ting Hu, Ronghang Hu, Chaitanya Ryali, Tengyu Ma, Haitham Khedr, Roman Rädle, Chloe Rolland, Laura Gustafson, et al. Sam 2: Segment anything in images and videos. arXiv preprint arXiv:2408.00714, 2024

    2024 arXiv

  40. [49]

    Grounded sam: Assembling open-world models for diverse visual tasks

    Tianhe Ren, Shilong Liu, Ailing Zeng, Jing Lin, Kunchang Li, He Cao, Jiayu Chen, Xinyu Huang, Yukang Chen, Feng Yan, et al. Grounded sam: Assembling open-world models for diverse visual tasks. arXiv preprint arXiv:2401.14159 , 2024

    2024 arXiv

  41. [50]

    Schwing �, and Oliver Wang �

    Zhongzheng Ren, Aseem Agarwala �, Bryan Russell �, Alexander G. Schwing �, and Oliver Wang �. Neural volu- metric object selection. In Conference on Computer Vision and Pattern Recognition (CVPR), 2022

    2022

  42. [51]

    Mask3D: Mask Trans- former for 3D Semantic Instance Segmentation

    Jonas Schult, Francis Engelmann, Alexander Hermans, Or Litany, Siyu Tang, and Bastian Leibe. Mask3D: Mask Trans- former for 3D Semantic Instance Segmentation. In Inter- national Conference on Robotics and Automation (ICRA) , 2023

    2023

  43. [52]

    Flashsplat: 2d to 3d gaussian splatting segmentation solved optimally

    Qiuhong Shen, Xingyi Yang, and Xinchao Wang. Flashsplat: 2d to 3d gaussian splatting segmentation solved optimally. In European Conference on Computer Vision (ECCV), 2025. 10

    2025

  44. [53]

    Deep marching tetrahedra: a hybrid represen- tation for high-resolution 3d shape synthesis

    Tianchang Shen, Jun Gao, Kangxue Yin, Ming-Yu Liu, and Sanja Fidler. Deep marching tetrahedra: a hybrid represen- tation for high-resolution 3d shape synthesis. In Advances in Neural Information Processing Systems (NeurIPS), 2021

    2021

  45. [54]

    Panoptic lifting for 3d scene understanding with neural fields

    Yawar Siddiqui, Lorenzo Porzi, Samuel Rota Bulò, Nor- man Müller, Matthias Nießner, Angela Dai, and Peter Kontschieder. Panoptic lifting for 3d scene understanding with neural fields. In Conference on Computer Vision and Pattern Recognition (CVPR), 2023

    2023

  46. [55]

    Silva, Mahtab Dahaghin, Matteo Toso, and Alessio Del Bue

    Myrna C. Silva, Mahtab Dahaghin, Matteo Toso, and Alessio Del Bue. Contrastive gaussian clus- tering: Weakly supervised 3d scene segmentation. https://arxiv.org/abs/2404.12784, 2024

    2024 arXiv

  47. [56]

    Julian Straub, Thomas Whelan, Lingni Ma, Yufan Chen, Erik Wijmans, Simon Green, Jakob J. Engel, Raul Mur-Artal, Carl Ren, Shobhit Verma, Anton Clarkson, Mingfei Yan, Brian Budge, Yajie Yan, Xiaqing Pan, June Yon, Yuyang Zou, Kimberly Leon, Nigel Carter, Jesus Briales, Tyler Gi...

    1906 arXiv

  48. [57]

    Sumner, Marc Pollefeys, Federico Tombari, and Francis Engelmann

    Ayça Takmaz, Elisabetta Fedele, Robert W. Sumner, Marc Pollefeys, Federico Tombari, and Francis Engelmann. OpenMask3D: Open-V ocabulary 3D Instance Segmentation. In Advances in Neural Information Processing Systems (NeurIPS), 2023

    2023

  49. [58]

    Searching efficient 3d archi- tectures with sparse point-voxel convolution

    Haotian Tang, Zhijian Liu, Shengyu Zhao, Yujun Lin, Ji Lin, Hanrui Wang, and Song Han. Searching efficient 3d archi- tectures with sparse point-voxel convolution. In European Conference on Computer Vision (ECCV), 2020

    2020

  50. [59]

    Isbnet: a 3d point cloud instance segmentation network with instance- aware sampling and box-aware dynamic convolution

    Khoi Nguyen Tuan Duc Ngo, Binh-Son Hua. Isbnet: a 3d point cloud instance segmentation network with instance- aware sampling and box-aware dynamic convolution. In Conference on Computer Vision and Pattern Recognition (CVPR), 2023

    2023

  51. [60]

    Depth-aware cnn for rgb-d segmentation

    Weiyue Wang and Ulrich Neumann. Depth-aware cnn for rgb-d segmentation. In European Conference on Computer Vision (ECCV), 2018

    2018

  52. [61]

    Opengaussian: Towards point-level 3d gaussian-based open vocabulary understand- ing

    Yanmin Wu, Jiarui Meng, Haijie Li, Chenming Wu, Yahao Shi, Xinhua Cheng, Chen Zhao, Haocheng Feng, Errui Ding, Jingdong Wang, and Jian Zhang. Opengaussian: Towards point-level 3d gaussian-based open vocabulary understand- ing. In Advances in Neural Information Processing Syste...

    2024

  53. [62]

    Blendedmvs: A large- scale dataset for generalized multi-view stereo networks

    Yao Yao, Zixin Luo, Shiwei Li, Jingyang Zhang, Yufan Ren, Lei Zhou, Tian Fang, and Long Quan. Blendedmvs: A large- scale dataset for generalized multi-view stereo networks. In Conference on Computer Vision and Pattern Recognition (CVPR), 2020

    2020

  54. [63]

    Gaussian grouping: Segment and edit anything in 3d scenes

    Mingqiao Ye, Martin Danelljan, Fisher Yu, and Lei Ke. Gaussian grouping: Segment and edit anything in 3d scenes. In European Conference on Computer Vision (ECCV), 2024

    2024

  55. [64]

    Scannet++: A high-fidelity dataset of 3d indoor scenes

    Chandan Yeshwanth, Yueh-Cheng Liu, Matthias Nießner, and Angela Dai. Scannet++: A high-fidelity dataset of 3d indoor scenes. In International Conference on Computer Vi- sion (ICCV), 2023

    2023

  56. [65]

    Omniseg3d: Omniversal 3d seg- mentation via hierarchical contrastive learning

    Haiyang Ying, Yixuan Yin, Jinzhi Zhang, Fan Wang, Tao Yu, Ruqi Huang, and Lu Fang. Omniseg3d: Omniversal 3d seg- mentation via hierarchical contrastive learning. In Confer- ence on Computer Vision and Pattern Recognition (CVPR) , 2024

    2024

  57. [66]

    Metascenes: Towards automated replica creation for real-world 3d scans

    Huangyue Yu, Baoxiong Jia, Yixin Chen, Yandan Yang, Puhao Li, Rongpeng Su, Jiaxin Li, Qing Li, Wei Liang, Zhu Song-Chun, Tengyu Liu, and Siyuan Huang. Metascenes: Towards automated replica creation for real-world 3d scans. In Conference on Computer Vision and Pattern Recogniti...

    2025

  58. [67]

    Manigaussian++: General robotic bimanual manipulation with hierarchical gaussian world model

    Tengbo Yu, Guanxing Lu, Zaijia Yang, Haoyuan Deng, Sea- son Si Chen, Jiwen Lu, Wenbo Ding, Guoqiang Hu, Yan- song Tang, and Ziwei Wang. Manigaussian++: General robotic bimanual manipulation with hierarchical gaussian world model. arXiv preprint arXiv:2506.19842, 2025

    2025 arXiv

  59. [68]

    Panop- ticrecon: Leverage open-vocabulary instance segmenta- tion for zero-shot panoptic reconstruction

    Xuan Yu, Yili Liu, Chenrui Han, Sitong Mao, Shunbo Zhou, Rong Xiong, Yiyi Liao, and Yue Wang. Panop- ticrecon: Leverage open-vocabulary instance segmenta- tion for zero-shot panoptic reconstruction. arXiv preprint arXiv:2407.01349, 2024

    2024 arXiv

  60. [69]

    Monoinstance: Enhancing monocular priors via multi-view instance align- ment for neural rendering and reconstruction

    Wenyuan Zhang, Yixiao Yang, Han Huang, Liang Han, Kanle Shi, Yu-Shen Liu, and Zhizhong Han. Monoinstance: Enhancing monocular priors via multi-view instance align- ment for neural rendering and reconstruction. In Conference on Computer Vision and Pattern Recognition (CVPR), 2025

    2025

  61. [70]

    Point transformer

    Hengshuang Zhao, Li Jiang, Jiaya Jia, Philip HS Torr, and Vladlen Koltun. Point transformer. In Conference on Com- puter Vision and Pattern Recognition (CVPR), 2021

    2021

  62. [71]

    In-place scene labelling and understanding with implicit scene representation

    Shuaifeng Zhi, Tristan Laidlow, Stefan Leutenegger, and An- drew J Davison. In-place scene labelling and understanding with implicit scene representation. In International Confer- ence on Computer Vision (ICCV), 2021

    2021

  63. [72]

    Feature 3dgs: Supercharg- ing 3d gaussian splatting to enable distilled feature fields

    Shijie Zhou, Haoran Chang, Sicheng Jiang, Zhiwen Fan, Ze- hao Zhu, Dejia Xu, Pradyumna Chari, Suya You, Zhangyang Wang, and Achuta Kadambi. Feature 3dgs: Supercharg- ing 3d gaussian splatting to enable distilled feature fields. In Conference on Computer Vision and Pattern Reco...

    2024

  64. [73]

    3d-vista: Pre-trained transformer for 3d vision and text alignment

    Zhu Ziyu, Ma Xiaojian, Chen Yixin, Deng Zhidong, Huang Siyuan, and Li Qing. 3d-vista: Pre-trained transformer for 3d vision and text alignment. In ICCV, 2023. 11 Table S.6. The selected id lists used for 3D object extraction experiment in Replica. Scene ID list office0 3,4,7,9...

    2023

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.