Pith. sign in

REVIEW 3 major objections 5 minor 49 references

QuerySplat: Decoupling Geometry and Appearance Representations in 3DGS Prediction

T0 review · 3 major / 5 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 QuerySplat separates geometry from appearance in a query decoder and reaches top DL3DV novel-view PSNR without input poses.

desk verdict Strong engineering contribution with an honest ablation story; the Appendix E sample filter and missing code/variance are the real reasons to hold before trusting the exact SOTA numbers. read the letter →

arxiv 2608.01186 v1 pith:LU7REEQR submitted 2026-08-02 cs.CV

Yinglong Li , Donghui Shen , Xiaoyu Zhang , Zhichao Ye , Hongyu Wu , Aimin Hao , Guofeng Zhang , Haomin Liu This is my paper
classification cs.CV
keywords 3DGaussianSplattingfeed-forwardreconstructionnovelviewsynthesispose-freequery-baseddecodergeometry-appearancedecouplingvisiongeometricmodelDL3DV
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

QuerySplat claims that the blurriness and geometric disorder of earlier query-based feed-forward 3D Gaussian Splatting come from forcing one set of queries to predict both spatial and appearance attributes at once. It builds a dual-branch decoder: a frozen pretrained vision geometric model supplies geometry features, cameras, and depth for a geometry branch, while a separate appearance branch attends to image and ray features to predict color and opacity. On DL3DV, the paper reports average PSNR gains of 2.30 dB over the best pose-free baseline and 1.04 dB over the best pose-required baseline, with sharper textures and more coherent Gaussian structure. A careful reader would care because it suggests sparse-view, pose-free 3D reconstruction from casual photos can be both fast and high-fidelity, without needing per-scene optimization or external camera calibration.

What carries the argument

The key mechanism is the attribute-aware dual-query decoder: two transformer decoder stacks with separate learnable queries. Geometry features come from a frozen VGGT-Ω (fused from several intermediate layers); appearance features combine RGB patch embeddings with Plücker ray embeddings computed from VGM-predicted cameras. The geometry branch emits positions, scales, and rotations as scene-level slots; the appearance branch reuses those geometry tokens as base state and adds queries for opacity and spherical-harmonic color. The procedure is completed by self-calibrated coordinate frames—two VGM passes aligned by Sim(3)—and transient early-stage regularization (bidirectional Chamfer distance

What would settle it

Compare QuerySplat against the same model fed oracle ground-truth poses on DL3DV: if the PSNR/SSIM gap is large, VGM camera error, not the decoder design, drives the reported gains. Separately, disable the early Chamfer/opacity regularizers on a set of scenes where VGM depth is known to be biased (mirrors, reflective floors, low-texture walls): if quality does not degrade, the depth-guided cold start is not the operative mechanism. Either observation would distinguish the decoupling hypothesis from the prior-quality hypothesis.

Watch

Extended reading notes

Core claim

The central claim is that Gaussian attributes have different information needs, and a query-based 3DGS predictor should decouple them. Geometry queries attend only to VGM geometry features and decode centers, scales, and rotations; appearance queries start from the decoded geometry tokens, add learnable queries, attend to RGB patch embeddings and Plücker rays, and decode opacities and spherical-harmonic colors. The VGM is frozen and defines a self-calibrated coordinate system, with an input-only pass for reconstruction and an all-view pass for supervision cameras aligned by Sim(3); an early Chamfer-distance and opacity-floor regularizer pulls the first Gaussian configurations toward the VGM

Load-bearing premise

The load-bearing premise is that the frozen VGM's predicted cameras, geometry features, and depth are accurate and consistent enough to define the coordinate system and the early training anchor; if its depth is biased on some scene type, the Chamfer cold-start could misdirect the Gaussian centers, and the pose-free claim inherits whatever camera error the VGM has.

Editorial extensions

If this is right

  • If correct, query-based non-pixel-aligned 3DGS is no longer inherently blurry: appearance-specific queries recover high-frequency detail while geometry queries keep scene coherence.
  • Pose-free reconstruction becomes a default property of the framework rather than an added module, since VGM-predicted cameras and coordinates replace external pose input.
  • The decoupling recipe transfers across VGM backbones (VGGT vs VGGT-Ω give comparable results), suggesting geometric-prior consumption, not a specific encoder, drives the benefit.
  • Optional test-time optimization in feature space provides an accuracy/runtime trade-off that improves even extrapolation views, not just input-view memorization.
  • Applications like repairing a 3D scene from video or reconstructing from text-to-video frames become feasible because the same feed-forward model lifts restored or generated 2D frames back into coherent 3DGS.

Reading between the lines

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

  • Beyond the paper's DL3DV experiments, the same dual-branch logic could be tested on other feed-forward representations (triplanes, voxels, mesh generators) where geometry and appearance also have different frequency and locality demands.
  • The paper's own Appendix A limits the fixed query budget for very large scenes; a reader might infer that region-splitting and merging is the next scaling step, but it is not implemented here.
  • Because early regularization leans on VGM depth, one can predict that scene classes where VGM depth is systematically biased (mirrors, transparent objects, repetitive texture) will cluster the failures; the paper does not report per-class breakdowns.
  • The pose-free comparison is clean only among native pose-free methods; as the appendix concedes, adapting a pose-required baseline like TokenGS with external VGGT-Ω cameras is not like-for-like, so the 1.04 dB posed-baseline gap should be read with that caveat.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. QuerySplat proposes a feed-forward 3D Gaussian Splatting predictor built around a dual-branch query decoder. Geometry queries attend to features from a frozen Vision Geometric Model (VGGT-Omega) and predict positions, scales, and rotations; appearance queries attend to RGB and Plücker ray embeddings and predict opacities and spherical-harmonic colors. The VGM also supplies camera poses and a Sim(3)-aligned coordinate system, which the paper argues makes the method pose-free. A Chamfer-distance and opacity-floor regularizer is applied only in early training to stabilize initialization. The model is trained on DL3DV and evaluated on DL3DV-Evaluation, reporting state-of-the-art interpolation results: average PSNR gains of 2.30 dB over the best pose-free baseline and 1.04 dB over the best pose-required baseline, with optional test-time optimization providing further gains.

Significance. If the claims hold, this is a meaningful advance: it shows that decoupling geometry and appearance in query-based 3DGS, combined with a frozen VGM prior, can match or exceed pixel-aligned feed-forward methods while remaining pose-free and fast. The controlled ablations in Table 2 (one-branch fusion drops PSNR by 1.98 dB, removing appearance by 4.25 dB) and the backbone swap in Table 4 (VGGT vs. VGGT-Omega within 0.18 dB) are genuine strengths and provide evidence that the architectural separation, not just the VGM, matters. However, the headline SOTA claim is weakened by an unexamined training-time sample filter that removes the 5% highest-loss training samples, and by the fact that all architectural ablations are run at a much smaller scale than the final model. These issues need to be resolved before the quantitative claims can be considered established.

major comments (3)
  1. [Appendix E / Figure 13] Appendix E states: 'During late training, a loss-rank filtering strategy retains 95% of samples according to their historical reconstruction losses.' This discards the 5% of training samples with the highest historical loss, i.e., the hardest cases. Since evaluation on DL3DV-Evaluation is not filtered, the model may be trained on an easy subset while tested on the full distribution, potentially inflating the headline average PSNR gains in Table 1. Figure 13's caption explicitly attributes an abrupt loss decrease to this filter, showing it materially alters training dynamics. There is no ablation of the retention rate, no characterization of discarded samples, and no comparison to training on all data. Because the central SOTA claim (2.30 dB over best pose-free, 1.04 dB over best pose-required) is an unfiltered evaluation average, this is a load-bearing unexamined choice; please add a no-
  2. [Ablation Studies / Appendix D] All architecture ablations (Tables 2-4) are trained only for 150K base iterations with 1,024 queries and no progressive query scaling, while the final model uses a 300K base stage plus progressive finetuning to 8,192 queries and the loss-rank filter. The paper argues the relative ordering is stable, but the magnitudes of the architecture benefits (e.g., 1.98 dB and 4.25 dB in Table 2) are not shown to persist in the final configuration, and the effect of progressive query scaling is not isolated. Since the headline comparison is the full pipeline, please demonstrate that the decoupling benefit remains at or near the final training protocol, at minimum without the loss-rank filter.
  3. [Self-Calibrated Coordinate System / Eq. (6)] The pose-free property and the early-stage regularization both rest on the accuracy and consistency of the frozen VGM. The Sim(3) alignment between the input-only and all-view VGM passes is assumed drift-free, and the Chamfer target in Eq. (6) is a point cloud back-projected from VGM depth. No pose or depth error analysis is reported (e.g., rotation/translation error against COLMAP on DL3DV-Evaluation), and no robustness test is provided. If VGM poses or depths are biased on some scene types, the coordinate alignment and the cold-start regularization could misdirect training. Please add a quantitative evaluation of VGM pose/depth accuracy and a sensitivity study with perturbed poses or depths.
minor comments (5)
  1. [Figure 13 caption] The parenthetical '(Note that each global step contains 500 (0.5k) steps with totally 150k steps)' is confusing and should be clarified or removed.
  2. [Appendix B] The in-the-wild TokenGS comparison concedes that 'scene sizes are not strictly consistent' due to TokenGS's rectangular input requirement. Please state precisely how this affects the qualitative comparison and whether the shown results are cropped or resized differently.
  3. [Tables 5-13] The paper reports averages over 300 cases per split without error bars or confidence intervals. Given the headline SOTA claim, please report standard errors or significance tests, especially for the 1.04 dB pose-required margin.
  4. [Appendix A] The limitations paragraph mentions the fixed query budget but does not quantify its effect on very large or complex scenes. A brief quantitative assessment (e.g., PSNR vs. scene extent) would be useful.
  5. [Notation] The VGM backbone is referred to as 'VGGT-Ω' in the text and 'VGGT-Omega' in some places; normalize the notation.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: QuerySplat's headline gains are external benchmark results, not reductions of fitted inputs.

full rationale

The central claim is an empirical comparison on DL3DV-Evaluation against external baselines run with official checkpoints. The geometry prior comes from a frozen, externally pretrained VGGT-Omega encoder; cameras, depth, and features are taken from that external model and are not fitted to the evaluation metrics. The predicted Gaussians are supervised by photometric rendering losses (L1/SSIM/LPIPS) on held-out views, so the reported 2.30 dB / 1.04 dB gains are not determined by construction. The early-stage Chamfer/opacity regularizers are transient (annealed over the first 20K iterations) and are shown by ablation to contribute only a small margin (22.90 vs 22.61 PSNR); this is a training initialization, not a fitted input renamed as a prediction. Appendix E's loss-rank filter ('During late training, a loss-rank filtering strategy retains 95% of samples according to their historical reconstruction losses...') is a training-sample-selection heuristic; it is unexamined and could affect generalization, but it does not define the evaluation quantity in terms of itself and is therefore a potential confound, not circularity. The few overlapping-author citations (PGSR, TokenSplat) appear only as related-work background and are not used to justify the paper's architecture or results. The pose-free property is inherited from the external VGM by design, not established by a self-referential argument. The stated limitation (fixed query budget) is a capacity constraint, not a circular step. Hence no load-bearing step reduces to its own inputs.

Assumptions & free parameters 6 free parameters · 4 assumptions · 0 invented entities

No new physical or conceptual entities are postulated; 'geometry queries' and 'appearance queries' are standard learnable network parameters, not entities with independent evidence requirements.

free parameters (6)
  • num_queries (1024 base, 8192 finetune) = 1024 / 8192
    Hand-chosen capacity budget; progressive doubling schedule is an ad hoc design choice affecting quality and runtime.
  • gaussians_per_query = 64
    Hand-chosen; sets primitive count per scene slot.
  • loss weights = lambda_ssim=0.2, lambda_lpips=0.05, lambda_vis=1.0, beta_cd=1.0, beta_alpha=0.1
    Hand-tuned training hyperparameters that directly affect reported metrics.
  • opacity floor alpha_min = 0.1
    Hand-chosen floor for early opacity regularization.
  • scale cap = 0.075
    Hand-chosen cap on activated Gaussian scales.
  • dataset-level translation-scale coefficient = estimated from 300 DL3DV cases
    Fitted constant used only for the non-like-for-like TokenGS in-the-wild comparison; not part of the main method.
assumptions (4)
  • domain assumption The frozen VGM (VGGT-Omega) encodes accurate multi-view geometry, cameras, and depth for the target scene distribution.
    Invoked throughout Method: geometry features F_geo, VGM-predicted cameras for supervision, and depth for Chamfer loss (Eq. 6). If VGM priors are wrong on a distribution, the pose-free claim and early regularization fail.
  • domain assumption Sim(3) alignment between two independent VGM passes is well-defined and does not introduce systematic drift.
    Self-Calibrated Coordinate System paragraph; relies on shared input views to recover the similarity transform.
  • standard math Differentiable 3D Gaussian splatting provides gradients sufficient to supervise scene-level, non-pixel-aligned parameters.
    Used in Loss Functions; standard renderer assumption of the field.
  • ad hoc to paper Loss-rank filtering that drops 5% of hard training samples does not bias learned generalization.
    Appendix E 'loss-rank filtering strategy retains 95% of samples'; an unexamined training trick that could inflate average metrics.

how reviews work

0 comments
Cite this review

Pith. "Pith review of QuerySplat: Decoupling Geometry and Appearance Representations in 3DGS Prediction." pith.science (2026). https://pith.science/paper/LU7REEQR

@misc{pith2026260801186,
  author       = {Pith},
  title        = {Pith review of: QuerySplat: Decoupling Geometry and Appearance Representations in 3DGS Prediction},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/LU7REEQR}},
  note         = {Machine review of arXiv:2608.01186}
}
read the original abstract

While feed-forward 3D Gaussian Splatting (3DGS) enables efficient 3D reconstruction, achieving high-fidelity rendering remains challenging. Existing pixel-aligned approaches suffer from spatial inflexibility and massive structural redundancy, whereas query-based methods lack 3D priors and entangle geometry with appearance, yielding blurry, pose-dependent results. To overcome these deficiencies, we propose \textbf{QuerySplat}, a feed-forward 3DGS framework driven by geometric priors and explicit appearance decoupling. Specifically, we design a dual-branch query-based decoder: the geometry branch leverages a pretrained Vision Geometric Model for spatial understanding, which intrinsically endows QuerySplat with pose-free modeling capabilities, while the appearance branch recovers high-frequency details through a dedicated pathway separated from geometric attribute regression. Extensive experiments demonstrate that QuerySplat mitigates the blurry rendering issues of early query-based models and consistently outperforms pixel-aligned approaches in rendering fidelity. On the challenging DL3DV benchmark, it achieves state-of-the-art novel view synthesis performance, with average PSNR gains of 2.30 dB and 1.04 dB over the best pose-free and pose-required baselines, respectively. Project Page: https://inspatio.github.io/querysplat.

Figures

Figures reproduced from arXiv: 2608.01186 by the authors.

Figure 1
Figure 1. QuerySplat overview. Given uncalibrated images with varying view counts, QuerySplat reconstructs clean and high-fidelity 3D Gaussian scenes in a pose-free, feed-forward, and non-pixel-aligned manner. Compared with prior methods, it produces substantially more coherent scene structure while supporting in-the-wild inputs within seconds. tiveness of the attribute-decoupled query mechanism. The main contributions are su… view at source ↗
Figure 2
Figure 2. Method Overview. A Vision Geometric Model (VGM) encodes geometry-aware memory and defines cameras. Geometry queries decode spatial Gaussian parameters from VGM features, while appearance queries read RGB/Plücker memory to predict the others. The resulting Gaussians are supervised by differentiable splatting in the VGM-defined coordinate system. Loss Functions The predicted Gaussians are rendered by differentiable Ga… view at source ↗
Figure 3
Figure 3. Qualitative comparison. QuerySplat preserves finer textures and sharper object boundaries, producing more faithful and visually detailed renderings than prior methods. Input AnySplat SplatWeaver YoNoSplat Ours [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (8 more)
Figure 4
Figure 4. Figure 4: In-the-wild qualitative comparison. Each row compares 3DGS reconstructions from casually captured unposed images. QuerySplat yields cleaner geometry and sharper details than competing pose-free methods. are initialized from pretrained ones with small perturbations q ne…
Figure 5
Figure 5. Figure 5: Feature aggregation variants. Illustration of (a) our dual-branch design, (b) one-branch feature fusion, and (c) geometry-only prediction without appearance features. and extrapolation views using PSNR, SSIM, and LPIPS. For fair comparison, inputs are resized to a shor…
Figure 9
Figure 9. Figure 9: 3D scene repair. Left: the manually corrupted Gaussian scene. Right: the reconstruction produced by QuerySplat after rendering the corrupted scene to video and applying video restoration. Synthesized Images Output 3DGS [PITH_FULL_IMAGE:figures/full_fig_p007_9.png]
Figure 10
Figure 10. Figure 10: 3D reconstruction from T2V-generated views. QuerySplat reconstructs a 3DGS scene from multi-view frames generated by a text-to-video model. The background is removed from the displayed 3DGS for visualization. ible everyday 3D capture. 3D Scene Repair. QuerySplat provi…
Figure 8
Figure 8. Figure 8: Geometric comparison. Left: point cloud obtained by back-projecting the depth predicted by VGGT-Ω. Right: centers of the 3D Gaussians predicted by QuerySplat. in the main model. Replacing the default backbone with VGGT (Wang et al. 2025a) yields comparable quality with…
Figure 11
Figure 11. Figure 11: Additional qualitative comparisons. We compare QuerySplat with representative posed and pose-free feed-forward 3DGS methods on diverse scenes. The red and blue boxes highlight enlarged regions. QuerySplat preserves sharper boundaries, finer textures, and more coherent…
Figure 12
Figure 12. Figure 12: Additional in-the-wild comparison with To￾kenGS. TokenGS is a pose-required method and therefore receives cameras from an external VGGT-Ω predictor, fol￾lowed by calibration to the DL3DV camera distribution used during its training. QuerySplat instead estimates camera…
Figure 13
Figure 13. Figure 13: Training-loss curves of the ablation variants. In the later training stage, the curves correspond from top to bottom to w/o Appearance, One-branch Query, w/o Early Reg., and the full model. All variants exhibit similar convergence trends and maintain a stable relative…

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

49 extracted references · 18 canonical work pages

  1. [1]

    Communication, Simulation, and Intelligent Agents: Implications of Personal Intelligent Machines for Medical Education

    Clancey, William J. Communication, Simulation, and Intelligent Agents: Implications of Personal Intelligent Machines for Medical Education. Proceedings of the Eighth International Joint Conference on Artificial Intelligence (IJCAI-83)

  2. [2]

    Classification Problem Solving

    Clancey, William J. Classification Problem Solving. Proceedings of the Fourth National Conference on Artificial Intelligence

  3. [3]

    , title =

    Robinson, Arthur L. , title =. 1980 , doi =. https://science.sciencemag.org/content/208/4447/1019.full.pdf , journal =

    1980

  4. [4]

    New Ways to Make Microcircuits Smaller---Duplicate Entry

    Robinson, Arthur L. New Ways to Make Microcircuits Smaller---Duplicate Entry. Science

  5. [5]

    Clancey and Glenn Rennels , abstract =

    Diane Warner Hasling and William J. Clancey and Glenn Rennels , abstract =. Strategic explanations for a diagnostic consultation system , journal =. 1984 , issn =. doi:https://doi.org/10.1016/S0020-7373(84)80003-6 , url =

    doi:10.1016/s0020-7373(84)80003-6 1984

  6. [6]

    and Rennels, Glenn R

    Hasling, Diane Warner and Clancey, William J. and Rennels, Glenn R. and Test, Thomas. Strategic Explanations in Consultation---Duplicate. The International Journal of Man-Machine Studies

  7. [7]

    Poligon: A System for Parallel Problem Solving

    Rice, James. Poligon: A System for Parallel Problem Solving

  8. [8]

    Transfer of Rule-Based Expertise through a Tutorial Dialogue

    Clancey, William J. Transfer of Rule-Based Expertise through a Tutorial Dialogue

Show all 49 references
  1. [9]

    The Engineering of Qualitative Models

    Clancey, William J. The Engineering of Qualitative Models

  2. [10]

    2023 , eprint=

    Attention Is All You Need , author=. 2023 , eprint=

    2023

  3. [11]

    Pluto: The 'Other' Red Planet

    NASA. Pluto: The 'Other' Red Planet

  4. [12]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , year =

    Jianyuan Wang and Minghao Chen and Shangzhan Zhang and Nikita Karaev and Johannes Sch. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , year =

  5. [13]

    CVPR , year=

    VGGT-Det: Mining VGGT Internal Priors for Sensor-Geometry-Free Multi-View Indoor 3D Object Detection , author=. CVPR , year=

  6. [14]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    TokenGS: Decoupling 3d gaussian prediction from pixels with learnable tokens , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  7. [15]

    Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

    No pose at all: Self-supervised pose-free 3d gaussian splatting from sparse views , author=. Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

  8. [16]

    Proceedings of the Computer Vision and Pattern Recognition Conference , pages=

    Depthsplat: Connecting gaussian splatting and depth , author=. Proceedings of the Computer Vision and Pattern Recognition Conference , pages=

  9. [17]

    arXiv preprint arXiv:2511.07321 , year=

    YoNoSplat: You Only Need One Model for Feedforward 3D Gaussian Splatting , author=. arXiv preprint arXiv:2511.07321 , year=

  10. [18]

    ACM Transactions on Graphics (TOG) , volume=

    Anysplat: Feed-forward 3d gaussian splatting from unconstrained views , author=. ACM Transactions on Graphics (TOG) , volume=. 2025 , publisher=

    2025

  11. [19]

    International Conference on Learning Representations , volume=

    No pose, no problem: Surprisingly simple 3d gaussian splats from sparse unposed images , author=. International Conference on Learning Representations , volume=

  12. [20]

    arXiv preprint arXiv:2605.07287 , year=

    SplatWeaver: Learning to Allocate Gaussian Primitives for Generalizable Novel View Synthesis , author=. arXiv preprint arXiv:2605.07287 , year=

    arXiv

  13. [21]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    DL3DV-10k: A large-scale scene dataset for deep learning-based 3d vision , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  14. [22]

    ACM Transactions on Graphics , volume=

    Stereo Magnification: Learning View Synthesis Using Multiplane Images , author=. ACM Transactions on Graphics , volume=. 2018 , publisher=

    2018

  15. [23]

    International Conference on Learning Representations , year=

    Decoupled Weight Decay Regularization , author=. International Conference on Learning Representations , year=

  16. [24]

    Proceedings of the Computer Vision and Pattern Recognition Conference , pages=

    VGGT: Visual geometry grounded transformer , author=. Proceedings of the Computer Vision and Pattern Recognition Conference , pages=

  17. [25]

    Proceedings of the Special Interest Group on Computer Graphics and Interactive Techniques Conference , pages=

    ArtiFixer: Enhancing and Extending 3D Reconstruction with Auto-Regressive Diffusion Models , author=. Proceedings of the Special Interest Group on Computer Graphics and Interactive Techniques Conference , pages=

  18. [26]

    Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

    pixelsplat: 3d gaussian splats from image pairs for scalable generalizable 3d reconstruction , author=. Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

  19. [27]

    Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

    Splatter image: Ultra-fast single-view 3d reconstruction , author=. Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

  20. [28]

    European conference on computer vision , pages=

    MVSplat: Efficient 3d gaussian splatting from sparse multi-view images , author=. European conference on computer vision , pages=. 2024 , organization=

    2024

  21. [29]

    , author=

    3d gaussian splatting for real-time radiance field rendering. , author=. ACM Trans. Graph. , volume=

  22. [30]

    IEEE Transactions on Visualization and Computer Graphics , volume=

    Pgsr: Planar-based gaussian splatting for efficient and high-fidelity surface reconstruction , author=. IEEE Transactions on Visualization and Computer Graphics , volume=. 2024 , publisher=

    2024

  23. [31]

    ACM Transactions on Graphics (ToG) , volume=

    Gaussian opacity fields: Efficient adaptive surface reconstruction in unbounded scenes , author=. ACM Transactions on Graphics (ToG) , volume=. 2024 , publisher=

    2024

  24. [32]

    European conference on computer vision , pages=

    End-to-end object detection with transformers , author=. European conference on computer vision , pages=. 2020 , organization=

    2020

  25. [33]

    Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

    Voxformer: Sparse voxel transformer for camera-based 3d semantic scene completion , author=. Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

  26. [34]

    Conference on robot learning , pages=

    Detr3d: 3d object detection from multi-view images via 3d-to-2d queries , author=. Conference on robot learning , pages=. 2022 , organization=

    2022

  27. [35]

    2023 IEEE International Conference on Robotics and Automation (ICRA) , pages=

    Mask3D: Mask transformer for 3d semantic instance segmentation , author=. 2023 IEEE International Conference on Robotics and Automation (ICRA) , pages=. 2023 , organization=

    2023

  28. [36]

    International Conference on Learning Representations , volume=

    Lrm: Large reconstruction model for single image to 3d , author=. International Conference on Learning Representations , volume=

  29. [37]

    Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

    DUST3R: Geometric 3d vision made easy , author=. Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

  30. [38]

    European conference on computer vision , pages=

    Grounding image matching in 3d with MAST3R , author=. European conference on computer vision , pages=. 2024 , organization=

    2024

  31. [39]

    Wang, Yifan and Zhou, Jianjun and Zhu, Haoyi and Chang, Wenzheng and Zhou, Yang and Li, Zizun and Chen, Junyi and Pang, Jiangmiao and Shen, Chunhua and He, Tong , journal=

  32. [40]

    arXiv preprint arXiv:2511.10647 , year=

    Depth Anything 3: Recovering the visual space from any views , author=. arXiv preprint arXiv:2511.10647 , year=

    arXiv

  33. [41]

    arXiv preprint arXiv:2604.08542 , year=

    Scal3r: Scalable test-time training for large-scale 3d reconstruction , author=. arXiv preprint arXiv:2604.08542 , year=

    arXiv

  34. [42]

    arXiv preprint arXiv:2603.04385 , year=

    Zipmap: Linear-time stateful 3d reconstruction via test-time training , author=. arXiv preprint arXiv:2603.04385 , year=

    arXiv

  35. [43]

    arXiv preprint arXiv:2604.14141 , year=

    Geometric Context Transformer for Streaming 3D Reconstruction , author=. arXiv preprint arXiv:2604.14141 , year=

    arXiv

  36. [44]

    arXiv preprint arXiv:2603.00697 , year=

    TokenSplat: Token-aligned 3d gaussian splatting for feed-forward pose-free reconstruction , author=. arXiv preprint arXiv:2603.00697 , year=

  37. [45]

    arXiv preprint arXiv:2512.10685 , year=

    Sharp monocular view synthesis in less than a second , author=. arXiv preprint arXiv:2512.10685 , year=

  38. [46]

    arXiv preprint arXiv:2412.09573 , year=

    FreeSplatter: Pose-free gaussian splatting for sparse-view 3d reconstruction , author=. arXiv preprint arXiv:2412.09573 , year=

    arXiv

  39. [47]

    arXiv preprint arXiv:2604.28193 , year=

    Generalizable Sparse-View 3D Reconstruction from Unconstrained Images , author=. arXiv preprint arXiv:2604.28193 , year=

    arXiv

  40. [48]

    arXiv preprint arXiv:2512.15508 , year=

    Off The Grid: Detection of Primitives for Feed-Forward 3D Gaussian Splatting , author=. arXiv preprint arXiv:2512.15508 , year=

  41. [49]

    arXiv preprint arXiv:2605.26115 , year=

    TriSplat: Simulation-ready feed-forward 3d scene reconstruction , author=. arXiv preprint arXiv:2605.26115 , year=

    arXiv

Pith tools

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