Pith. sign in

REVIEW 4 major objections 6 minor 31 references

CrossView-GS: Cross-view Gaussian Splatting For Large-scale Scene Reconstruction

T0 review · 4 major / 6 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read A multi-branch Gaussian splatting pipeline overcomes the gradient smoothing that limits cross-view scene reconstruction, delivering higher-quality novel views than prior state-of-the-art methods on aerial-ground, pure-aerial, and…

desk verdict A solid cross-view 3DGS method with a clear gradient-smoothing diagnosis, but the 'superior performance' claim is undercut by the paper's own LPIPS numbers. read the letter →

arxiv 2501.01695 v2 pith:S24XWUJH submitted 2025-01-03 cs.CV

Chenhao Zhang , Yuanping Cao , Lei Zhang This is my paper · ORCID
classification cs.CV
keywords 3DGaussianSplattingcross-viewreconstructionnovelviewsynthesislarge-scalesceneadaptivedensificationcontrolmulti-branchfusiongradient-awareregularization
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 claims that the standard densification rule in 3D Gaussian Splatting fails when training images come from sharply different viewpoints, such as aerial and ground cameras, because averaging gradients across such views washes out the salient gradients that tell the model where to add Gaussian primitives. To fix this, the authors train separate sub-models on each view set, use the distant-view sub-model to initialize the cross-view model, replace the average-gradient densification test with a per-view maximum-gradient test, and add a regularization term that penalizes regions where the cross-view model renders worse than the branch models. They then fuse Gaussians that each branch created but the cross-view model lacks, judged by a voxel-grid overlap test, and fine-tune briefly. The reported result is consistently higher novel-view quality than prior methods on aerial-ground, pure-aerial, and pure-ground benchmarks, with the largest gains in aerial views and roughly 1.15 dB average PSNR improvement over the Scaffold-GS baseline.

What carries the argument

The load-bearing mechanism is the max-gradient densification rule of Eq. (4), which replaces the average-gradient test of the original 3DGS densification with a maximum over per-view gradient statistics, so a primitive is densified if any single view set finds it salient. Around this rule the paper builds a three-part pipeline: multi-branch initialization, where sub-models are trained separately on each view set and the distant-view sub-model is downsampled into the starting point cloud; gradient-aware regularization, where branch renderings serve as pseudo-labels in a $\max(0, d(\mathrm{pred},\mathrm{gt})-d(\mathrm{ref},\mathrm{gt}))$ loss (Eq. 5) that only penalizes the cross-view model where it is worse than a branch; and unique Gaussian supplementation (Eq. 6), which keeps branch Gaussians whose anchor voxel is absent from the cross-view model and then fine-tunes. The voxel grid shared by the Scaffold-GS models is the object that lets the supplementation test decide which Gaussians are complementary.

What would settle it

A direct test would be to train a cross-view scene twice: once with the stated supplementation and once after applying a small rigid perturbation (for example a 10 cm translation) to the branch model coordinates before the voxel-overlap test of Eq. (6), while keeping the rest of the pipeline identical. If the perturbed version matches or exceeds the PSNR of the unperturbed version on held-out aerial or ground views, then the precise voxel alignment is not what drives the gains, and the supplementation mechanism would need a different explanation.

Watch

Extended reading notes

Core claim

The paper's central claim is that cross-view Gaussian splatting can be made to work by treating each view set as an independent teacher instead of letting all views compete inside one optimization. The authors show that when views with large disparities are pooled, the densification condition $\frac{\nabla_1+\nabla_2}{c_1+c_2}>\tau$ (Eq. 3) is dominated by the larger, smoother-view gradient statistics, so primitives with salient per-view gradients are never densified. Their remedy, the max-gradient selection $\max(\nabla_1/c_1,\nabla_2/c_2)>\tau$ (Eq. 4), restores single-view-level densification, and the pseudo-label regularization (Eq. 5) pushes the cross-view model to be at least as good as the best branch. Finally, a voxel-grid membership test (Eq. 6) adds branch-only Gaussians to the cross-view model so fine details from each viewpoint set are retained. The paper argues that this combination yields superior novel-view synthesis, with the largest quality gains on aerial views.

Load-bearing premise

The method assumes that the branch sub-models and the cross-view model use identical, aligned voxel grids with the same coordinate frame, so the test $\nu_i \notin V_f$ reliably identifies Gaussians that are genuinely complementary; if the grids are even slightly misaligned, the supplementation step can insert redundant or misplaced Gaussians and the reported gains may not transfer.

Editorial extensions

If this is right

  • On the tested benchmarks, the method reports higher PSNR, SSIM, and lower LPIPS than 3DGS, Mip-Splatting, Scaffold-GS, Hier-GS, and Horizon-GS, with an average 1.15 dB PSNR gain over Scaffold-GS.
  • Aerial views benefit most from the gradient-aware regularization and the distant-view initialization, making the method well suited to drone-plus-ground captures.
  • The same pipeline transfers to pure-aerial and pure-ground pairs with strong view differences, and to three view sets, indicating the approach is not limited to aerial-ground data.
  • The cross-view model trains in 50k iterations (30k initial plus 20k fine-tuning), far fewer than UC-GS's 900k, while staying competitive or better on the UC-GS benchmark.

Reading between the lines

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

  • The per-view maximum-gradient test in Eq. (4) could be dropped into any 3DGS-style densification routine, not just Scaffold-GS, whenever training views are clustered by appearance or viewpoint; the paper only demonstrates it inside its own pipeline.
  • The multi-branch pseudo-label loss is effectively a knowledge-distillation signal from view-specific teachers to the fused model, so the same regularization may apply to other cross-domain radiance-field tasks such as day-night or indoor-outdoor reconstruction.
  • The voxel-alignment assumption suggests a testable boundary: on scenes with noisy or drifting pose estimates, the supplementation step may degrade unless the branch and fused models are re-registered before Eq. (6); enforcing alignment explicitly could make the method more robust.
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 / 6 minor

Summary. The paper proposes CrossView-GS, a 3D Gaussian Splatting method for reconstructing large-scale scenes from cross-view image sets, such as aerial and ground views. The method has three components: (1) multi-branch initialization, where sub-models are trained on separate view sets and the distant-view sub-model's Gaussians are downsampled to produce the initial point cloud for a cross-view model; (2) gradient-aware regularization, which modifies the densification criterion to use the maximum per-view gradient (Eq. 4) and adds a pseudo-label loss that penalizes predictions worse than the branch-model renderings (Eq. 5); and (3) unique Gaussian supplementation, which inserts Gaussians from branch models whose anchors lie outside the cross-view model's voxel grid (Eq. 6) and fine-tunes the combined model. Experiments on MatrixCity, UC-GS, ISPRS, and two newly collected internet-video datasets report PSNR, SSIM, and LPIPS against 3DGS, Mip-Splatting, Scaffold-GS, Hier-GS, and Horizon-GS.

Significance. The motivation is sound: cross-view data violate the smooth-view assumption of adaptive densification control in 3DGS, and the paper gives a clear, arithmetically correct derivation of why per-view salient gradients get averaged away (Eqs. 2-4). The proposed max-gradient rule is a simple, plausible fix. The systematic ablation in Table V demonstrates that each component contributes to PSNR/SSIM/LPIPS gains over the Scaffold-GS baseline, and the inclusion of real captured (internet) datasets and a limitations section (dynamic objects) are positives. However, the unqualified claim of 'superior performance' in the abstract and conclusion is undercut by the paper's own LPIPS numbers in Table I, where CrossView-GS is worse than Horizon-GS on three of four aerial-ground scenes. No error bars or multi-seed statistics are reported, and the evaluation protocol for the UC-GS-derived scenes is incompletely specified. These issues must be resolved before the central claim can be accepted.

major comments (4)
  1. [Abstract; Table I] The abstract and conclusion claim 'superior performance' in novel view synthesis compared to state-of-the-art methods. However, Table I shows that CrossView-GS has higher (worse) LPIPS than Horizon-GS on Block small (0.213 vs. 0.197), NYC (0.199 vs. 0.189), and SF (0.269 vs. 0.240), and it ties with Horizon-GS on SSIM on Block small (0.846) while being worse on SF SSIM (0.784 vs. 0.791). Since LPIPS is a standard perceptual metric reported by the paper itself, the unqualified superiority claim is not supported. Please either qualify the claim (e.g., 'better PSNR/SSIM, comparable or worse LPIPS'), or provide multi-seed runs with error bars and a statistical test demonstrating that the aggregate improvement is significant despite these regressions.
  2. [Section V-A, Table I] The quantitative comparison for the UC-GS-derived NYC and SF scenes in Table I uses a newly introduced split (one test image out of every eight based on image suffix number). The paper does not state whether the baseline results in Table I were obtained by retraining each baseline on this new split or were copied from the original papers. If baselines were not retrained on this split, the comparison is unfair and the reported margins may not be meaningful. Please specify the exact split, describe how each baseline was adapted to it, and, ideally, release the train/test index files for reproducibility.
  3. [Section IV-C, Eq. (6)] The unique Gaussian supplementation step defines non-overlap as νi ∉ Vf , where Vf is the voxel grid of the cross-view model and νi is the anchor voxel of a Gaussian from a branch model. This test is only valid if the branch-model grids and the cross-view model grid share the same coordinate frame, voxel origin, and lattice. The paper only states that the models are 'reconstructed using Scaffold-GS with the same voxel size.' Since Scaffold-GS anchors are created during training and depend on the initial point cloud and densification history, and the branch models and cross-view model are initialized from different point clouds, the grids are not guaranteed to be aligned. If they are misaligned, the supplementation will insert redundant or displaced Gaussians and the method's reported gains may not transfer to other scenes. Please specify the alignment procedure or provide evidence (e.g., a quantitative analysis of grid overlap) that the grids are co-registered.
  4. [Section IV-B, Eq. (5)] The regularization loss Lreg uses renderings from the multi-branch sub-models as pseudo-labels for the training views. Because these sub-models are trained on the same training images that are later used to train the cross-view model, this is a self-distillation mechanism. If a sub-model's rendering is erroneous on a given view (e.g., due to incomplete geometry or view-specific artifacts), the loss penalizes the cross-view model for deviating from that error, potentially reinforcing it. The paper does not discuss this failure mode or provide evidence that the pseudo-labels are reliable, such as an ablation that replaces the pseudo-label with the ground truth or a comparison of pseudo-label quality on held-out views. Please add such an analysis or justify why the risk is negligible.
minor comments (6)
  1. [Section V, Baselines] The statement 'All GS-based methods are initialized with the same point cloud' is ambiguous: does this mean the baselines also use the point cloud derived from the distant-view sub-model, or that all methods use the same COLMAP point cloud? This is important because the proposed initialization alone yields a large gain (Table V, 'w/ Init.'). Please clarify.
  2. [Table IV] The table caption notes that the SOTA methods are trained for 900k iterations while the proposed method uses 30k+20k. This is a strength, but the iteration counts should be stated directly in the table or caption so readers can verify the comparison without reading the text.
  3. [Figure 3] The caption says 'variation of maximum gradient in different views' but the axes are not labeled. Please label the axes and clarify what is plotted (e.g., maximum gradient over iterations).
  4. [References] References [5] and [6] are incompletely specified: [5] is a bare arXiv identifier with no year/venue, and [6] is an arXiv preprint; please complete the bibliographic information for these entries.
  5. [Section V-D] The limitations section mentions dynamic objects but does not discuss the effect of the newly introduced split or the lack of code; consider adding a reproducibility statement.
  6. [Appendix A] The dataset collection for Temple and Monument relies on MASt3R-SfM for pose estimation; please report the pose accuracy or registration errors for these datasets, since the cross-view alignment is critical to the method.

Circularity Check

0 steps flagged · score 2.0 of 10

No significant circularity: the central derivation is self-contained; only a minor non-load-bearing self-citation appears.

full rationale

The paper's load-bearing components — multi-branch initialization, max-gradient densification (Eq. 4), pseudo-label regularization (Eq. 5), and unique Gaussian supplementation (Eq. 6) — are all defined operationally from the training views and Scaffold-GS internals, not from the test outputs. The pseudo-label ref in Eq. (5) is a rendering of sub-models trained on the same training views, so it is a self-distillation regularizer rather than a fitted parameter renamed as a prediction; it can propagate errors but does not make the reported novel-view results true by construction. No result is imported from a same-author uniqueness theorem, and the only overlapping-author reference ([2], a survey co-authored by Yuanping Cao) is introductory and not load-bearing. The claimed 'superior performance' is in tension with the LPIPS results in Table I (CrossView-GS is worse than Horizon-GS on Block small, NYC, and SF), but that is a correctness or evidence concern, not circularity, because the evaluation is on held-out views and no metric is defined in terms of the method's own outputs. Score 2 reflects the minor self-citation and the mild self-distillation flavor, not a circular derivation.

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

The central claim depends on four hyperparameters and an unreported voxel size, plus assumptions about 3DGS densification behavior and the reliability of branch models. No new physical entities are postulated. The free parameters are all chosen empirically, not derived.

free parameters (6)
  • λreg = 1.0
    Weight of the gradient-aware regularization loss in Eq. (8); chosen manually, not derived.
  • λSSIM = 0.2
    Weight of SSIM loss following Scaffold-GS; standard setting.
  • λvol = 0.01
    Weight of volume regularization following Scaffold-GS; standard setting.
  • τ (downsampling ratio) = 10
    Ratio used to downsample Gaussian centers from distant view branch to create initial point cloud (Eq. 1); chosen by hand.
  • training iterations = 30k + 20k
    Iterations for cross-view model and fine-tuning; chosen empirically, not justified by convergence analysis.
  • voxel size = not specified
    The voxel size for overlap judgment in unique Gaussian supplementation is not reported, yet the method's correctness depends on it.
assumptions (4)
  • standard math The gradient of the union view set is a weighted average of the per-view gradients, so its maximum can be smaller than the larger per-view gradient.
    Used in Section IV-B Eq. (3) to motivate the max-gradient rule; a standard property of weighted averages.
  • domain assumption Scaffold-GS's adaptive densification relies on average gradients over all training views, and improving gradient selection improves reconstruction.
    The paper modifies the criterion inside Scaffold-GS; this assumes the bottleneck is gradient statistics.
  • ad hoc to paper Multi-branch sub-models trained on individual view sets provide reliable pseudo-labels for regularization.
    The loss in Eq. (5) assumes the branch renderings are better references than ground truth when the cross-view model is worse; this is not verified independently.
  • domain assumption The voxel grids of the branch models and cross-view model are aligned and comparable, allowing overlap determination.
    Section IV-C uses voxel membership to decide which Gaussians are unique; the paper states the same voxel size is used but does not describe how alignment is ensured.

how reviews work

0 comments
Cite this review

Pith. "Pith review of CrossView-GS: Cross-view Gaussian Splatting For Large-scale Scene Reconstruction." pith.science (2026). https://pith.science/paper/S24XWUJH

@misc{pith2026250101695,
  author       = {Pith},
  title        = {Pith review of: CrossView-GS: Cross-view Gaussian Splatting For Large-scale Scene Reconstruction},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/S24XWUJH}},
  note         = {Machine review of arXiv:2501.01695}
}
read the original abstract

3D Gaussian Splatting (3DGS) leverages densely distributed Gaussian primitives for high-quality scene representation and reconstruction. While existing 3DGS methods perform well in scenes with minor view variation, large view changes from cross-view data pose optimization challenges for these methods. To address these issues, we propose a novel cross-view Gaussian Splatting method for large-scale scene reconstruction based on multi-branch construction and fusion. Our method independently reconstructs models from different sets of views as multiple independent branches to establish the baselines of Gaussian distribution, providing reliable priors for cross-view reconstruction during initialization and densification. Specifically, a gradient-aware regularization strategy is introduced to mitigate smoothing issues caused by significant view disparities. Additionally, a unique Gaussian supplementation strategy is utilized to incorporate complementary information of multi-branch into the cross-view model. Extensive experiments on benchmark datasets demonstrate that our method achieves superior performance in novel view synthesis compared to state-of-the-art methods.

Figures

Figures reproduced from arXiv: 2501.01695 by the authors.

Figure 1
Figure 1. Left: The cross-view data poses challenges for large-scale scene reconstruction due to significant view variation. Middle: Our method CrossView-GS, demonstrates the ability to reconstruct finer structures with more consistent geometry and appearance than Scaffold-GS [1] across diverse scenes. Right: The quantitative comparisons with additional SOTA methods highlight the superiority of our method in reconstructing sc… view at source ↗
Figure 2
Figure 2. An overview of the proposed method CrossView-GS. The method first initializes the multi-branch structure comprising several sub-models while establishing the cross-view model based on distant views. Then, a gradient-sensitive regularization strategy is used to reconstruct the cross-view model. Finally, unique Gaussian primitives are fused to supplement the cross-view model, achieving large-scale scene reconstruction… view at source ↗
Figure 3
Figure 3. The first row shows the variation of maximum gradient in different [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (9 more)
Figure 4
Figure 4. Figure 4: The rendering results and zoomed-in images obtained by using our CrossView-GS, as well as some SOTA methods like Hier-GS and Scaffold-GS. [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 5
Figure 5. Figure 5: Qualitative ablations. (a) Cross-view reconstruction using the baseline [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]
Figure 6
Figure 6. Figure 6: Limitation. The cross-view inconsistency caused by dynamic objects results in unnatural artifacts. VI. CONCLUSION We propose a novel method CrossView-GS, for large-scale scene reconstruction based on cross-view data. The multi- [PITH_FULL_IMAGE:figures/full_fig_p007_6.png]
Figure 7
Figure 7. Figure 7: More rendering results and zoomed-in images obtained by our method, as well as some SOTA methods in aerial-ground scenes. [PITH_FULL_IMAGE:figures/full_fig_p009_7.png]
Figure 8
Figure 8. Figure 8: More rendering results and zoomed-in images obtained by our method, as well as some SOTA methods in Monument with aerial views and Block [PITH_FULL_IMAGE:figures/full_fig_p010_8.png]
Figure 9
Figure 9. Figure 9: Point clouds and final reconstruction results by using different initializations. [PITH_FULL_IMAGE:figures/full_fig_p010_9.png]
Figure 10
Figure 10. Figure 10: Qualitative ablations. Column 1: The results of baseline in cross-view reconstruction. Column 2: The results of baseline with initialization from distant aerial views. Column 3: The results of baseline with gradient-aware regularization. Column 4: The results of basel…
Figure 11
Figure 11. Figure 11: Visualization of Temple and Monument, as well as example views [PITH_FULL_IMAGE:figures/full_fig_p011_11.png]
Figure 12
Figure 12. Figure 12: Visualization of unique Gaussian supplementation. We respectively [PITH_FULL_IMAGE:figures/full_fig_p011_12.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

31 extracted references · 26 canonical work pages

  1. [1]

    Scaffold- gs: Structured 3d gaussians for view-adaptive rendering,

    T. Lu, M. Yu, L. Xu, Y . Xiangli, L. Wang, D. Lin, and B. Dai, “Scaffold- gs: Structured 3d gaussians for view-adaptive rendering,” in Proceedings of CVPR , 2024

    work page 2024

  2. [2]

    Recent advances in 3d gaussian splatting,

    T. Wu, Y .-J. Yuan, L.-X. Zhang, J. Yang, Y .-P. Cao, L.-Q. Yan, and L. Gao, “Recent advances in 3d gaussian splatting,” Computational Visual Media, vol. 10, no. 4, pp. 613–642, 2024

    2024

  3. [3]

    3d gaussian splatting as new era: a survey,

    B. Fei, J. Xu, R. Zhang, Q. Zhou, W. Yang, and Y . He, “3d gaussian splatting as new era: a survey,” IEEE Transactions on Visualization and Computer Graphics , pp. 1–20, 2024

    work page 2024

  4. [4]

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

    B. Kerbl, G. Kopanas, T. Leimk ¨uhler, and G. Drettakis, “3d gaussian splatting for real-time radiance field rendering,” ACM Transactions on Graphics (TOG), vol. 42, no. 4, pp. 1–14, 2023

    work page 2023

  5. [5]

    Mvgs: Multi-view-regulated gaussian splatting for novel view synthesis,

    X. Du, Y . Wang, and X. Yu, “Mvgs: Multi-view-regulated gaussian splatting for novel view synthesis,” 2024

    work page 2024

  6. [6]

    Drone-assisted road gaussian splatting with cross-view uncertainty,

    S. Zhang, B. Ye, X. Chen, Y . Chen, Z. Zhang, C. Peng, Y . Shi, and H. Zhao, “Drone-assisted road gaussian splatting with cross-view uncertainty,” arXiv preprint arXiv:2408.15242 , 2024

    arXiv 2024

  7. [7]

    Dragon: Drone and ground gaussian splatting for 3d building reconstruction,

    Y . Ham, M. Michalkiewicz, and G. Balakrishnan, “Dragon: Drone and ground gaussian splatting for 3d building reconstruction,” inProceedings of ICIP , 2024

    work page 2024

  8. [8]

    Spatial-aware feature aggregation for image based cross-view geo-localization,

    Y . Shi, L. Liu, X. Yu, and H. Li, “Spatial-aware feature aggregation for image based cross-view geo-localization,” 2019

    work page 2019

Show all 31 references
  1. [9]

    Cvm-net: Cross-view matching network for image-based ground-to-aerial geo-localization,

    S. Hu, M. Feng, R. M. H. Nguyen, and G. H. Lee, “Cvm-net: Cross-view matching network for image-based ground-to-aerial geo-localization,” in Proceedings of CVPR , 2018

    2018

  2. [10]

    Bridging the domain gap for ground-to-aerial image matching,

    K. Regmi and M. Shah, “Bridging the domain gap for ground-to-aerial image matching,” in Proceedings of ICCV , 2019

    2019

  3. [11]

    Pixel-gs: Density control with pixel-aware gradient for 3d gaussian splatting,

    Z. Zhang, W. Hu, Y . Lao, T. He, and H. Zhao, “Pixel-gs: Density control with pixel-aware gradient for 3d gaussian splatting,” in Proceedings of ECCV, 2024

    2024

  4. [12]

    Revising densification in gaussian splatting,

    S. Rota Bulo, L. Porzi, and P. Kontschieder, “Revising densification in gaussian splatting,” in Proceedings of ECCV , 2024

    2024

  5. [13]

    Mip-splatting: Alias-free 3d gaussian splatting,

    Z. Yu, A. Chen, B. Huang, T. Sattler, and A. Geiger, “Mip-splatting: Alias-free 3d gaussian splatting,” in Proceedings of CVPR , 2024

    2024

  6. [14]

    Bungeenerf: Progressive neural radiance field for extreme multi- scale scene rendering,

    Y . Xiangli, L. Xu, X. Pan, N. Zhao, A. Rao, C. Theobalt, B. Dai, and D. Lin, “Bungeenerf: Progressive neural radiance field for extreme multi- scale scene rendering,” in Proceedings of ECCV , 2022

    2022

  7. [15]

    Citygaussian: Real-time high-quality large-scale scene rendering with gaussians,

    Y . Liu, C. Luo, L. Fan, N. Wang, J. Peng, and Z. Zhang, “Citygaussian: Real-time high-quality large-scale scene rendering with gaussians,” in Proceedings of ECCV , 2024

    2024

  8. [16]

    Vastgaussian: Vast 3d gaussians for large scene reconstruction,

    J. Lin, Z. Li, X. Tang, J. Liu, S. Liu, J. Liu, Y . Lu, X. Wu, S. Xu, Y . Yan, and W. Yang, “Vastgaussian: Vast 3d gaussians for large scene reconstruction,” in Proceedings of CVPR , 2024

    2024

  9. [17]

    A hierarchical 3d gaussian representation for real-time rendering of very large datasets,

    B. Kerbl, A. Meuleman, G. Kopanas, M. Wimmer, A. Lanvin, and G. Drettakis, “A hierarchical 3d gaussian representation for real-time rendering of very large datasets,” ACM Transactions on Graphics (TOG), vol. 43, no. 4, pp. 1–15, 2024

    2024

  10. [18]

    Cross-view image geolocalization,

    T.-Y . Lin, S. Belongie, and J. Hays, “Cross-view image geolocalization,” in Proceedings of CVPR , 2013

    2013

  11. [19]

    Ground and aerial meta- data integration for localization and reconstruction: A review,

    X. Gao, S. Shen, Z. Hu, and Z. Wang, “Ground and aerial meta- data integration for localization and reconstruction: A review,” Pattern Recognition Letters , vol. 127, pp. 202–214, 2019

    2019

  12. [20]

    3d reconstruction through fusion of cross-view images,

    R. Qin, S. Song, X. Ling, and M. Elhashash, “3d reconstruction through fusion of cross-view images,” Recent Advances in Image Restoration with Applications to Real World Problems , p. 123, 2020

    2020

  13. [21]

    Leveraging photogrammetric mesh models for aerial-ground feature point matching toward integrated 3d reconstruction,

    Q. Zhu, Z. Wang, H. Hu, L. Xie, X. Ge, and Y . Zhang, “Leveraging photogrammetric mesh models for aerial-ground feature point matching toward integrated 3d reconstruction,” ISPRS Journal of Photogrammetry and Remote Sensing , vol. 166, pp. 26–40, 2020

    2020

  14. [22]

    Dust3r: Geometric 3d vision made easy,

    S. Wang, V . Leroy, Y . Cabon, B. Chidlovskii, and J. Revaud, “Dust3r: Geometric 3d vision made easy,” in Proceedings of CVPR , 2024

    2024

  15. [23]

    Horizon-gs: Unified 3d gaussian splatting for large-scale aerial- to-ground scenes,

    L. Jiang, K. Ren, M. Yu, L. Xu, J. Dong, T. Lu, F. Zhao, D. Lin, and B. Dai, “Horizon-gs: Unified 3d gaussian splatting for large-scale aerial- to-ground scenes,” arXiv preprint arXiv:2412.01745 , 2024

    2024 arXiv

  16. [24]

    Pygs: Large-scale scene representation with pyramidal 3d gaussian splatting,

    Z. Wang and D. Xu, “Pygs: Large-scale scene representation with pyramidal 3d gaussian splatting,” 2024

    2024

  17. [25]

    Facenet: A unified em- bedding for face recognition and clustering,

    F. Schroff, D. Kalenichenko, and J. Philbin, “Facenet: A unified em- bedding for face recognition and clustering,” in Proceedings of CVPR , 2015

    2015

  18. [26]

    Matrixcity: A large-scale city dataset for city-scale neural rendering and beyond,

    Y . Li, L. Jiang, L. Xu, Y . Xiangli, Z. Wang, D. Lin, and B. Dai, “Matrixcity: A large-scale city dataset for city-scale neural rendering and beyond,” in Proceedings of ICCV , 2023

    2023

  19. [27]

    Isprs benchmark for multi-platform photogrammetry,

    F. Nex, M. Gerke, F. Remondino, H.-J. Przybilla, M. B ¨aumker, and A. Zurhorst, “Isprs benchmark for multi-platform photogrammetry,” ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences , vol. 2, pp. 135–142, 2015

    2015

  20. [28]

    Image quality assessment: from error visibility to structural similarity,

    Z. Wang, A. C. Bovik, H. R. Sheikh, and E. P. Simoncelli, “Image quality assessment: from error visibility to structural similarity,” IEEE transactions on image processing , vol. 13, no. 4, pp. 600–612, 2004

    2004

  21. [29]

    Mixture of volumetric primitives for efficient neural render- ing,

    S. Lombardi, T. Simon, G. Schwartz, M. Zollhoefer, Y . Sheikh, and J. Saragih, “Mixture of volumetric primitives for efficient neural render- ing,” ACM Transactions on Graphics (ToG) , vol. 40, no. 4, pp. 1–13, 2021

    2021

  22. [30]

    The unreasonable effectiveness of deep features as a perceptual metric,

    R. Zhang, P. Isola, A. A. Efros, E. Shechtman, and O. Wang, “The unreasonable effectiveness of deep features as a perceptual metric,” in Proceedings of CVPR , 2018, pp. 586–595

    2018

  23. [31]

    Mast3r-sfm: a fully-integrated solution for unconstrained structure-from-motion,

    B. Duisterhof, L. Zust, P. Weinzaepfel, V . Leroy, Y . Cabon, and J. Revaud, “Mast3r-sfm: a fully-integrated solution for unconstrained structure-from-motion,” arXiv preprint arXiv:2409.19152 , 2024. 9 SF Ground truthCrossView-GS (Ours)Hier-GSScaffold-GS Block_small Zeche Bloc...

    2024 arXiv

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.