Pith. sign in

REVIEW 4 major objections 5 minor 1 cited by

SLGaussian: Fast Language Gaussian Splatting in Sparse Views

T0 review · 4 major / 5 minor · reviewed 2026-08-11 · deepseek-v4-flash

Pith's one-line read SLGaussian claims the first feed-forward construction of a language-embedded 3D Gaussian scene from two views, replacing per-scene optimization with a frozen geometry predictor and a learned label-lifting branch.

desk verdict A genuinely new feed-forward semantic field pipeline with a real speed story, but the empirical case needs cleanup before the headline claims can be taken at face value. read the letter →

arxiv 2412.08331 v3 pith:DHBLWSFN submitted 2024-12-11 cs.CV

classification cs.CV
keywords 3DGaussianSplattingsemanticfieldopen-vocabularyquerysparse-viewreconstructionfeed-forwardinferenceCLIPmaskassociationlanguagememorybank
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

SLGaussian claims that a 3D semantic field can be predicted in a single forward pass from just two RGB views, rather than optimized per scene. The authors propose a two-branch network: a frozen feed-forward Gaussian predictor supplies geometry and appearance, while a learned branch turns consistent 2D segmentations into per-Gaussian 3D label-ID features. A multi-view language memory bank binds those labels to CLIP embeddings so arbitrary text can query the scene. If the claim is right, language-queryable 3D reconstruction becomes practical for robots or AR devices with very few cameras: no per-scene retraining, scene inference in roughly 25–30 seconds, and open-vocabulary queries in 0.011 seconds each. On LERF and 3D-OVS, the reported metrics are substantially better than the LERF and LangSplat baselines under the same two-view protocol.

What carries the argument

The central object is the semantic Gaussian, written $\{o_i,\mu_i,\delta_i,f_i\}$: the first three parameters come from a frozen feed-forward Gaussian predictor and carry geometry and appearance, while $f_i$ is a 3D label-ID feature predicted by a CNN-plus-MLP branch from multi-view semantic label maps. Semantic splatting renders these features with the standard Gaussian rasterization equation, so every pixel receives a label-ID vector rather than a color. The other load-bearing mechanism is the multi-view language memory bank, which assigns evenly spaced vectors in $[0,1]^3$ as IDs, stores the two-view CLIP embeddings for each object under that ID, and answers queries by mapping rendered IDs back to CLIP space and scoring against the text embedding. Two auxiliary moves stabilize the pipeline: input views are duplicated five times and tracked with a video object segmentation model so that SAM masks agree across sparse views, and the duplicated tracks are majority-voted per pixel before training.

What would settle it

Run the frozen model on two-view pairs from a held-out indoor dataset where the geometry branch is known to reconstruct well, and query an object category that never appeared in training; if the relevance heatmap is spatially incoherent or the chosen IoU falls to chance while the RGB render is sharp, the label-lifting branch has not generalized.

Watch

Extended reading notes

Core claim

The paper's central claim is that consistent 2D mask labels plus CLIP embeddings can be transferred into 3D Gaussian semantic parameters in a single forward pass, because a frozen feed-forward Gaussian reconstruction supplies the geometry. The semantic branch predicts a 3D feature $f_i$ per Gaussian, which is rendered by the same splatting equation used for color; the rendered feature map is then matched, through low-dimensional label IDs stored in a multi-view language memory bank, to the CLIP encoding of an arbitrary query text. This replaces the per-scene autoencoder compression used by earlier language-3D methods with a fixed ID-to-CLIP lookup, avoiding memory overflow and the neutralization of inconsistent multi-view language features. The authors report that this design yields sparse-view open-vocabulary localization and segmentation with two views, and reduces scene-inference and query latency by roughly an order of magnitude relative to the main baseline.

Load-bearing premise

The entire pipeline rests on the transfer of the label-lifting branch from its training setup—YouTube video frames at 256×256 with video-tracked segmentations—to completely different scenes, unseen object categories, and 416×576 resolution, with no RGB-based correction of predicted labels.

Editorial extensions

If this is right

  • A two-view input with poses is enough to construct a queryable 3D semantic field; no per-scene optimization or retraining is needed on a new scene.
  • Open-vocabulary queries become cheap enough for interactive use: roughly 0.011 seconds per query, versus seconds for pipelines that re-segment and re-encode 2D views.
  • The geometry branch is interchangeable, so any feed-forward Gaussian predictor that improves sparse-view geometry should immediately improve the semantic field built on top of it.
  • Because label IDs stay constant while CLIP features are stored per object, the same scene supports many query phrasings without re-rendering the semantic field.
  • Mask association via repeated-frame video tracking is what makes SAM usable under extreme sparsity, so the consistency problem is solved at the supervision level rather than by dense multi-view optimization.

Reading between the lines

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

  • One consequence the paper leaves implicit: the semantic labels are only as good as the frozen geometry branch, since a Gaussian that the geometry predictor fails to create can never receive a semantic label.
  • A natural stress test is to compare query IoU on objects whose RGB reconstruction is sharp versus objects that are thin, reflective, or textureless, isolating where the label-lifting fails.
  • Another extension is to replace CLIP features in the memory bank with other 2D vision-language embeddings; if the speedup survives, the label-ID indexing is a general mechanism rather than a CLIP-specific fix.
  • A further test would train the semantic branch at the inference resolution of 416×576 rather than at 256×256, isolating how much of the reported gain comes from the label-ID design and how much from resolution shift.
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 / 5 minor

Summary. The paper proposes SLGaussian, a feed-forward method that constructs a 3D semantic field from two RGB views without per-scene optimization. The pipeline uses MVSplat to predict base Gaussian parameters, SAM with video-object tracking (applied to five replicated copies of each view) to obtain multi-view consistent masks, and a lightweight CNN/MLP branch to predict per-Gaussian semantic label IDs. A multi-view language memory bank stores CLIP embeddings keyed by low-dimensional label IDs, and open-vocabulary queries are performed by rendering the semantic label map, translating IDs to CLIP features, and thresholding a relevance score against the query text. Experiments on LERF and 3D-OVS report improvements over LERF, LangSplat, and 3D-OVS in IoU, localization accuracy, and mIoU, along with claims of fast inference and 0.011-second queries.

Significance. If the central claims hold, SLGaussian would be a useful step toward sparse-view, feed-forward 3D semantic fields, with practical implications for robotics and AR/VR. The memory-bank design is a reasonable alternative to per-scene autoencoders, and the ablation in Table 3 indicates that each component contributes to the reported gains. However, the empirical support is currently incomplete: key comparison tables are partial, the query threshold is unspecified, and the timing claims are internally inconsistent. The paper does not release code or weights and reports no error bars, so the quantitative advantage cannot yet be independently verified.

major comments (4)
  1. [§4.1 vs. Abstract/Introduction] The timing claims are contradictory as written. Section 4.1 states that scene inference with 416×576 images takes approximately 0.3 seconds, while the abstract and introduction claim 'under 30 seconds' and 'around 25 seconds'. These may refer to different pipeline stages, but the paper never provides a decomposition. Because fast scene inference is a headline contribution, the authors should report the end-to-end time including SAM segmentation, video tracking, mask voting, CLIP encoding, and memory-bank construction, and reconcile the 0.3-second and 25-second figures.
  2. [Table 2, §4.4] Table 2 reports IoU and mIoU for only 5 of the 10 3D-OVS scenes and defers the remaining five to the supplementary material, yet the text claims that 'results across ten scenes' are shown and that the overall mIoU 'significantly surpasses' baselines. The overall row in Table 2 is not auditable from the paper alone, and the selection of which five scenes appear in the main table is not justified. All ten scenes should be reported in the main text, or the authors should state a pre-specified selection rule.
  3. [§3.3, Eq. (6)] The query threshold n is never specified. Section 3.3 says all pixels with relevance scores above a threshold n are selected as the query result, but there is no description of how n is chosen, whether it is fixed across scenes and queries, or whether it is tuned per scene. If n is tuned on the test scenes, the reported IoU numbers are not a fair comparison with baselines. The authors should specify a fixed threshold or a calibration protocol and report sensitivity to n.
  4. [§3.1, §4.1] The semantic branch is trained on RealEstate10K at 256×256 resolution and applied to LERF and 3D-OVS at 416×576 with object categories and viewpoint distributions not seen in training, without per-scene adaptation. The paper argues for generalization but provides only two scene-level ablations (Table 3) and no statistical summary across the full test sets. The authors should provide per-scene results for all evaluation scenes, error bars or multiple runs, and a discussion of how the resolution and domain gap affect the predicted semantic fields.
minor comments (5)
  1. [Introduction, Related Work] There are several typos, including '3DGS-baesd' and 'mathods', and the phrase 'we will seek a more efficient feature embedding method' reads like a proposal rather than a description of the completed method.
  2. [§4.4] The text says 'Results across ten scenes are shown in Figure 2', but Figure 2 is the method overview; the intended reference is likely Table 2.
  3. [§4.3, Table 1] The metric 'chosen IoU' is not defined. The authors should explain how query regions are selected and how 'chosen' differs from standard IoU, especially because the threshold n in Eq. (6) directly affects this metric.
  4. [Table 3] The first row of Table 3 appears to reproduce the LangSplat numbers from Table 1, but the configuration labeled 'without feed-forward model, without MA, without MV-LMB' is not clearly described. The authors should specify what model produces this row and how it relates to LangSplat.
  5. [§3.2] The view-duplication factor of five and the choice of a 3-dimensional label space are introduced without any sensitivity analysis; a short ablation of these hyperparameters would strengthen the paper.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the semantic field is lifted from external SAM/CLIP features via a frozen feed-forward model; no test-set fitting or self-citation chain forces the reported results.

full rationale

The paper's derivation chain is self-contained against external benchmarks and does not reduce by construction to its own inputs. The base Gaussians come from a frozen MVSplat feed-forward model, and the semantic branch takes SAM-plus-tracking mask maps as input and is trained on RealEstate10K with an MSE loss against mask identifiers; at test time on LERF and 3D-OVS the model is applied without per-scene optimization. The open-vocabulary querying uses a memory bank built from CLIP embeddings of the input views, but this is an explicit algorithmic component, not a hidden re-use of the evaluation labels. The reported IoU/localization numbers are compared against external baselines on external datasets, so no fitted parameter is renamed as a prediction. Self-citations such as LangSplat are used as baselines or in ablations, not as load-bearing justification for the method's central premise. The paper does have auditability gaps that are not circularity: Table 2 defers half of the 3D-OVS scenes to the supplementary, the threshold n in Eq. (6) is unspecified, and the reported inference time is inconsistent (0.3 s vs. 25 s). These are correctness and reproducibility concerns, not evidence that any claimed result is equivalent to its inputs by definition.

Assumptions & free parameters 5 free parameters · 5 assumptions · 0 invented entities

The central claim rests on several external components and design choices that are not derived or validated in the main text. The most load-bearing are: SAM and the video tracker providing consistent masks, CLIP features being view-invariant enough for the query aggregation in Eq. (6), the frozen MVSplat geometry being accurate on the test scenes, and the random 3D label IDs being learnable and transferable across datasets and resolutions. The free parameters (label dimension, duplication factor, threshold, canonical phrases) are chosen by hand and not ablated in the main paper.

free parameters (5)
  • Semantic label dimension = 3
    The 3D ID space [0,1]^3 is chosen by design; the paper states a 1D space is not enough to separate object labels (Section 3.2).
  • View duplication factor = 5
    Input views are replicated five times to make a 10-frame sequence for video tracking; the factor is a hand-chosen heuristic (Section 3.2).
  • Query relevance threshold n = not specified
    Pixels with relevance above n are selected as query results; the paper does not state how n is set (Section 3.3).
  • Canonical phrases = 'object', 'things', 'stuff', 'texture'
    Used as negative/phrase anchors in Eq (6); chosen without ablation in the main text (Section 3.3).
  • Semantic feature dimension for Gaussians = 3
    Same as the label ID dimension; f_i is 3-dimensional (Section 3.1).
assumptions (5)
  • domain assumption SAM masks provide consistent, accurate object-level segmentation of the relevant objects in sparse views.
    The whole pipeline uses SAM masks as the source of object labels; if SAM fails, the masks and memory bank are wrong. Invoked in Section 3.1 and 3.2.
  • domain assumption Video object tracking (Cutie) can align SAM masks across the duplicated 10-frame sequence.
    Eq (4) assumes the tracked frames produce consistent labels per pixel; the vote chooses the most frequent label.
  • domain assumption CLIP embeddings of the same object from different views are similar enough that the max/min aggregation in Eq (6) yields correct relevance scores.
    The query formula assumes at least one view's CLIP embedding of the object matches the query and the canonical phrases suppress background.
  • domain assumption The learned mapping from 2D label maps to per-Gaussian 3D labels trained on RealEstate10K at 256x256 transfers to LERF/3D-OVS at 416x576.
    Training resolution and data differ from evaluation; generalization is assumed and tested only on a subset of scenes in the main text.
  • domain assumption MVSplat predicts sufficiently accurate base Gaussians from two views.
    The semantic branch operates on Gaussians from the frozen MVSplat model, so poor geometry breaks the semantic field.

how reviews work

0 comments
Cite this review

Pith. "Pith review of SLGaussian: Fast Language Gaussian Splatting in Sparse Views." pith.science (2026). https://pith.science/paper/DHBLWSFN

@misc{pith2026241208331,
  author       = {Pith},
  title        = {Pith review of: SLGaussian: Fast Language Gaussian Splatting in Sparse Views},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/DHBLWSFN}},
  note         = {Machine review of arXiv:2412.08331}
}
read the original abstract

3D semantic field learning is crucial for applications like autonomous navigation, AR/VR, and robotics, where accurate comprehension of 3D scenes from limited viewpoints is essential. Existing methods struggle under sparse view conditions, relying on inefficient per-scene multi-view optimizations, which are impractical for many real-world tasks. To address this, we propose SLGaussian, a feed-forward method for constructing 3D semantic fields from sparse viewpoints, allowing direct inference of 3DGS-based scenes. By ensuring consistent SAM segmentations through video tracking and using low-dimensional indexing for high-dimensional CLIP features, SLGaussian efficiently embeds language information in 3D space, offering a robust solution for accurate 3D scene understanding under sparse view conditions. In experiments on two-view sparse 3D object querying and segmentation in the LERF and 3D-OVS datasets, SLGaussian outperforms existing methods in chosen IoU, Localization Accuracy, and mIoU. Moreover, our model achieves scene inference in under 30 seconds and open-vocabulary querying in just 0.011 seconds per query.

Figures

Figures reproduced from arXiv: 2412.08331 by the authors.

Figure 1
Figure 1. With just two RGB views, our method infers 3D semantic fields in under 30 seconds without per-scene optimization. [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Overview of our approach. Starting with two-view RGB images, we first apply SAM [ [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Visualization of RealEstate10K dataset scenes. The [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Qualitative comparisons of open-vocabulary 3D object localization on the LERF and 3D-OVS datasets. The top row [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]
Figure 5
Figure 5. Figure 5: Qualitative comparisons of open-vocabulary 3D object segmentation on the 3D-OVS dataset. [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

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

  1. SplatSearch: Instance Image Goal Navigation for Mobile Robots using 3D Gaussian Splatting and Diffusion Models

    cs.RO 2025-11 conditional novelty 6.0 of 10

    SplatSearch combines sparse-view 3D Gaussian Splatting, multi-view diffusion inpainting, and semantic/visual frontier scoring to achieve viewpoint-invariant instance image-goal navigation in unknown environments.

Reference graph

Works this paper leans on

65 extracted references · 17 canonical work pages · cited by 1 Pith paper

  1. [1]

    Jonathan T Barron, Ben Mildenhall, Matthew Tancik, Peter Hedman, Ricardo Martin-Brualla, and Pratul P Srinivasan. 2021. Mip-nerf: A multiscale repre- sentation for anti-aliasing neural radiance fields. In Proceedings of the IEEE/CVF international conference on computer vision . 5855–5864

  2. [2]

    Jiazhong Cen, Jiemin Fang, Chen Yang, Lingxi Xie, Xiaopeng Zhang, Wei Shen, and Qi Tian. 2023. Segment Any 3D Gaussians. arXiv preprint arXiv:2312.00860 (2023)

  3. [3]

    David Charatan, Sizhe Lester Li, Andrea Tagliasacchi, and Vincent Sitzmann

  4. [4]

    Anpei Chen, Zexiang Xu, Andreas Geiger, Jingyi Yu, and Hao Su. 2022. Tensorf: Tensorial radiance fields. In European conference on computer vision . Springer, 333–350

  5. [5]

    Boyuan Chen, Fei Xia, Brian Ichter, Kanishka Rao, Keerthana Gopalakrishnan, Michael S Ryoo, Austin Stone, and Daniel Kappler. 2023. Open-vocabulary queryable scene representations for real world planning. In 2023 IEEE Interna- tional Conference on Robotics and Automation (ICRA) . IEEE, 11509–11522

  6. [6]

    Yuedong Chen, Haofei Xu, Chuanxia Zheng, Bohan Zhuang, Marc Pollefeys, Andreas Geiger, Tat-Jen Cham, and Jianfei Cai. 2024. Mvsplat: Efficient 3d gaussian splatting from sparse multi-view images.arXiv preprint arXiv:2403.14627 (2024)

  7. [7]

    Yuedong Chen, Chuanxia Zheng, Haofei Xu, Bohan Zhuang, Andrea Vedaldi, Tat- Jen Cham, and Jianfei Cai. 2024. MVSplat360: Feed-Forward 360 Scene Synthesis from Sparse Views. arXiv preprint arXiv:2411.04924 (2024)

  8. [8]

    Ho Kei Cheng, Seoung Wug Oh, Brian Price, Alexander Schwing, and Joon-Young Lee. 2023. Tracking anything with decoupled video segmentation. In Proceedings of the IEEE/CVF International Conference on Computer Vision . 1316–1326

Show all 65 references
  1. [9]

    Runyu Ding, Jihan Yang, Chuhui Xue, Wenqing Zhang, Song Bai, and Xiaojuan Qi. 2023. Pla: Language-driven open-vocabulary 3d scene understanding. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition . 7010–7019

  2. [10]

    Bin Dou, Tianyu Zhang, Yongjia Ma, Zhaohui Wang, and Zejian Yuan. 2024. Cosseggaussians: Compact and swift scene segmenting 3d gaussians. arXiv preprint arXiv:2401.05925 (2024)

  3. [11]

    Golnaz Ghiasi, Xiuye Gu, Yin Cui, and Tsung-Yi Lin. 2022. Scaling open- vocabulary image segmentation with image-level labels. In European Conference on Computer Vision. Springer, 540–557

  4. [12]

    Rahul Goel, Dhawal Sirikonda, Saurabh Saini, and PJ Narayanan. 2023. Interactive segmentation of radiance fields. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition . 4201–4211

  5. [13]

    Qiao Gu, Zhaoyang Lv, Duncan Frost, Simon Green, Julian Straub, and Chris Sweeney. 2024. EgoLifter: Open-world 3D Segmentation for Egocentric Percep- tion. arXiv preprint arXiv:2403.18118 (2024)

  6. [14]

    Huy Ha and Shuran Song. 2022. Semantic abstraction: Open-world 3d scene understanding from 2d vision-language models. arXiv preprint arXiv:2207.11514 (2022)

  7. [15]

    Bing He, Yunuo Chen, Guo Lu, Li Song, and Wenjun Zhang. 2024. S4D: Streaming 4D Real-World Reconstruction with Gaussians and 3D Control Points. arXiv preprint arXiv:2408.13036 (2024)

  8. [16]

    Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. 2016. Deep residual learning for image recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition . 770–778

  9. [17]

    Chenguang Huang, Oier Mees, Andy Zeng, and Wolfram Burgard. 2023. Visual language maps for robot navigation. In 2023 IEEE International Conference on Robotics and Automation (ICRA). IEEE, 10608–10615

  10. [18]

    Krishna Murthy Jatavallabhula, Alihusein Kuwajerwala, Qiao Gu, Mohd Omama, Tao Chen, Alaa Maalouf, Shuang Li, Ganesh Iyer, Soroush Saryazdi, Nikhil Keetha, et al. 2023. Conceptfusion: Open-set multimodal 3d mapping. arXiv preprint arXiv:2302.07241 (2023)

  11. [19]

    Yuzhou Ji, He Zhu, Junshu Tang, Wuyi Liu, Zhizhong Zhang, Yuan Xie, Lizhuang Ma, and Xin Tan. 2024. FastLGS: Speeding up Language Embedded Gaussians with Feature Grid Mapping. arXiv preprint arXiv:2406.01916 (2024)

  12. [20]

    Bernhard Kerbl, Georgios Kopanas, Thomas Leimkühler, and George Drettakis

  13. [21]

    Justin Kerr, Chung Min Kim, Ken Goldberg, Angjoo Kanazawa, and Matthew Tancik. 2023. Lerf: Language embedded radiance fields. In Proceedings of the IEEE/CVF International Conference on Computer Vision . 19729–19739

  14. [22]

    Alexander Kirillov, Eric Mintun, Nikhila Ravi, Hanzi Mao, Chloe Rolland, Laura Gustafson, Tete Xiao, Spencer Whitehead, Alexander C Berg, Wan-Yen Lo, et al

  15. [23]

    Sosuke Kobayashi, Eiichi Matsumoto, and Vincent Sitzmann. 2022. Decomposing nerf for editing via feature field distillation. Advances in Neural Information Processing Systems 35 (2022), 23311–23330

  16. [24]

    Boyi Li, Kilian Q Weinberger, Serge Belongie, Vladlen Koltun, and René Ranftl

  17. [25]

    In Proceedings of the IEEE/CVF International Conference on Computer Vision

    Segment anything. In Proceedings of the IEEE/CVF International Conference on Computer Vision. 4015–4026

  18. [26]

    Wanhua Li, Yujie Zhao, Minghan Qin, Yang Liu, Yuanhao Cai, Chuang Gan, and Hanspeter Pfister. 2025. LangSplatV2: High-dimensional 3D Language Gaussian Splatting with 450+ FPS. arXiv preprint arXiv:2507.07136 (2025)

  19. [27]

    Wanhua Li, Renping Zhou, Jiawei Zhou, Yingwei Song, Johannes Herter, Minghan Qin, Gao Huang, and Hanspeter Pfister. 2025. 4d langsplat: 4d language gaussian splatting via multimodal large language models. In Proceedings of the Computer Vision and Pattern Recognition Conference...

  20. [28]

    Zhan Li, Zhang Chen, Zhong Li, and Yi Xu. 2024. Spacetime gaussian feature splatting for real-time dynamic view synthesis. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition . 8508–8520

  21. [29]

    Hao Li, Roy Qin, Zhengyu Zou, Diqi He, Bohan Li, Bingquan Dai, Dingewn Zhang, and Junwei Han. 2024. Langsurf: Language-embedded surface gaussians for 3d scene understanding. arXiv preprint arXiv:2412.17635 (2024)

  22. [30]

    Fangfu Liu, Wenqiang Sun, Hanyang Wang, Yikai Wang, Haowen Sun, Junliang Ye, Jun Zhang, and Yueqi Duan. 2024. ReconX: Reconstruct Any Scene from Sparse Views with Video Diffusion Model.arXiv preprint arXiv:2408.16767 (2024)

  23. [31]

    Kunhao Liu, Fangneng Zhan, Jiahui Zhang, Muyu Xu, Yingchen Yu, Abdulmotaleb El Saddik, Christian Theobalt, Eric Xing, and Shijian Lu. 2023. Weakly supervised 3d open-vocabulary segmentation. Advances in Neural Information Processing Systems 36 (2023), 53433–53456

  24. [32]

    Weijie Lyu, Xueting Li, Abhijit Kundu, Yi-Hsuan Tsai, and Ming-Hsuan Yang

  25. [33]

    Feng Liang, Bichen Wu, Xiaoliang Dai, Kunpeng Li, Yinan Zhao, Hang Zhang, Peizhao Zhang, Peter Vajda, and Diana Marculescu. 2023. Open-vocabulary semantic segmentation with mask-adapted clip. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition ...

  26. [34]

    Ben Mildenhall, Pratul P Srinivasan, Matthew Tancik, Jonathan T Barron, Ravi Ramamoorthi, and Ren Ng. 2021. Nerf: Representing scenes as neural radiance fields for view synthesis. Commun. ACM 65, 1 (2021), 99–106

  27. [35]

    Songyou Peng, Kyle Genova, Chiyu Jiang, Andrea Tagliasacchi, Marc Pollefeys, Thomas Funkhouser, et al. 2023. Openscene: 3d scene understanding with open vocabularies. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. 815–824

  28. [36]

    Minghan Qin, Wanhua Li, Jiawei Zhou, Haoqian Wang, and Hanspeter Pfister

  29. [37]

    arXiv preprint arXiv:2404.07977 (2024)

    Gaga: Group Any Gaussians via 3D-aware Memory Bank. arXiv preprint arXiv:2404.07977 (2024)

  30. [38]

    Kirill Mazur, Edgar Sucar, and Andrew J Davison. 2023. Feature-realistic neural fusion for real-time, open set scene understanding. In 2023 IEEE International Conference on Robotics and Automation (ICRA) . IEEE, 8201–8207

  31. [39]

    Ayça Takmaz, Elisabetta Fedele, Robert W Sumner, Marc Pollefeys, Federico Tombari, and Francis Engelmann. 2023. Openmask3d: Open-vocabulary 3d instance segmentation. arXiv preprint arXiv:2306.13631 (2023)

  32. [40]

    Jiaxiang Tang, Jiawei Ren, Hang Zhou, Ziwei Liu, and Gang Zeng. 2023. Dream- gaussian: Generative gaussian splatting for efficient 3d content creation. arXiv preprint arXiv:2309.16653 (2023)

  33. [41]

    Yunsong Wang, Tianxin Huang, Hanlin Chen, and Gim Hee Lee. 2024. FreeSplat: Generalizable 3D Gaussian Splatting Towards Free-View Synthesis of Indoor Scenes. arXiv preprint arXiv:2405.17958 (2024)

  34. [42]

    In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition

    Langsplat: 3d language gaussian splatting. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition . 20051–20060

  35. [43]

    Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, et al. 2021. Learning transferable visual models from natural language supervision. In International conference on machine learni...

  36. [44]

    Jin-Chuan Shi, Miao Wang, Hao-Bin Duan, and Shao-Hua Guan. 2024. Language embedded 3d gaussians for open-vocabulary scene understanding. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition . 5333– 5343

  37. [45]

    Haofei Xu, Songyou Peng, Fangjinhua Wang, Hermann Blum, Daniel Barath, Andreas Geiger, and Marc Pollefeys. 2024. Depthsplat: Connecting gaussian splatting and depth. arXiv preprint arXiv:2410.13862 (2024)

  38. [46]

    Ziyi Yang, Xinyu Gao, Wen Zhou, Shaohui Jiao, Yuqing Zhang, and Xiaogang Jin. 2024. Deformable 3d gaussians for high-fidelity monocular dynamic scene reconstruction. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 20331–20341

  39. [47]

    Zeyu Yang, Hongye Yang, Zijie Pan, Xiatian Zhu, and Li Zhang. 2023. Real-time photorealistic dynamic scene representation and rendering with 4d gaussian splatting. arXiv preprint arXiv:2310.10642 (2023)

  40. [48]

    Guanjun Wu, Taoran Yi, Jiemin Fang, Lingxi Xie, Xiaopeng Zhang, Wei Wei, Wenyu Liu, Qi Tian, and Xinggang Wang. 2024. 4d gaussian splatting for real-time dynamic scene rendering. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition . 20310–20320

  41. [49]

    Yanmin Wu, Jiarui Meng, Haijie Li, Chenming Wu, Yahao Shi, Xinhua Cheng, Chen Zhao, Haocheng Feng, Errui Ding, Jingdong Wang, et al. 2024. OpenGaus- sian: Towards Point-Level 3D Gaussian-based Open Vocabulary Understanding. arXiv preprint arXiv:2406.02058 (2024). MM ’25, Octob...

  42. [50]

    Haolin Xiong, Sairisheek Muttukuru, Rishi Upadhyay, Pradyumna Chari, and Achuta Kadambi. 2023. Sparsegs: Real-time 360{\deg} sparse view synthesis using gaussian splatting. arXiv preprint arXiv:2312.00206 (2023)

  43. [51]

    Alex Yu, Vickie Ye, Matthew Tancik, and Angjoo Kanazawa. 2021. pixelnerf: Neural radiance fields from one or few images. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition . 4578–4587

  44. [52]

    Chuanrui Zhang, Yingshuang Zou, Zhuoling Li, Minmin Yi, and Haoqian Wang

  45. [53]

    Dongbin Zhang, Yunfei Liu, Lijian Lin, Ye Zhu, Kangjie Chen, Minghan Qin, Yu Li, and Haoqian Wang. 2025. HRAvatar: High-Quality and Relightable Gauss- ian Head Avatar. In Proceedings of the Computer Vision and Pattern Recognition Conference. 26285–26296

  46. [54]

    Botao Ye, Sifei Liu, Haofei Xu, Xueting Li, Marc Pollefeys, Ming-Hsuan Yang, and Songyou Peng. 2024. No Pose, No Problem: Surprisingly Simple 3D Gaussian Splats from Sparse Unposed Images. arXiv preprint arXiv:2410.24207 (2024)

  47. [55]

    Mingqiao Ye, Martin Danelljan, Fisher Yu, and Lei Ke. 2023. Gaussian grouping: Segment and edit anything in 3d scenes. arXiv preprint arXiv:2312.00732 (2023)

  48. [56]

    Lin Yen-Chen, Pete Florence, Jonathan T Barron, Tsung-Yi Lin, Alberto Rodriguez, and Phillip Isola. 2022. Nerf-supervision: Learning dense object descriptors from neural radiance fields. In 2022 international conference on robotics and automation (ICRA). IEEE, 6496–6503

  49. [57]

    Zehao Zhu, Zhiwen Fan, Yifan Jiang, and Zhangyang Wang. 2023. Fsgs: Real-time few-shot view synthesis using gaussian splatting. arXiv preprint arXiv:2312.00451 (2023)

  50. [59]

    arXiv preprint arXiv:2408.13770 (2024)

    Transplat: Generalizable 3d gaussian splatting from sparse multi-view images with transformers. arXiv preprint arXiv:2408.13770 (2024)

  51. [61]

    Dongbin Zhang, Chuming Wang, Weitao Wang, Peihao Li, Minghan Qin, and Hao- qian Wang. 2024. Gaussian in the Wild: 3D Gaussian Splatting for Unconstrained Image Collections. arXiv preprint arXiv:2403.15704 (2024)

  52. [62]

    Shuaifeng Zhi, Tristan Laidlow, Stefan Leutenegger, and Andrew J. Davison. 2021. In-Place Scene Labelling and Understanding with Implicit Scene Representation

  53. [63]

    Tinghui Zhou, Richard Tucker, John Flynn, Graham Fyffe, and Noah Snavely

  54. [2018]

    arXiv preprint arXiv:1805.09817 (2018)

    Stereo magnification: Learning view synthesis using multiplane images. arXiv preprint arXiv:1805.09817 (2018)

  55. [2022]

    arXiv preprint arXiv:2201.03546 (2022)

    Language-driven semantic segmentation. arXiv preprint arXiv:2201.03546 (2022)

  56. [2023]

    ACM Transactions on Graphics 42, 4 (July 2023)

    3D Gaussian Splatting for Real-Time Radiance Field Rendering. ACM Transactions on Graphics 42, 4 (July 2023). https://repo-sam.inria.fr/fungraph/3d- gaussian-splatting/

  57. [2024]

    In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition

    pixelsplat: 3d gaussian splats from image pairs for scalable generalizable 3d reconstruction. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 19457–19467

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.