Pith. sign in

REVIEW 4 major objections 5 minor 47 references

Automated 3D-GS Registration and Fusion via Skeleton Alignment and Gaussian-Adaptive Features

T0 review · 4 major / 5 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read Independently built 3D-GS sub-maps can be aligned and fused automatically via skeleton alignment and covariance-aware features, cutting rotation error 41.9% and lifting PSNR by 10.11 dB.

desk verdict A plausible 3D-GS registration/fusion idea undercut by a garbled core equation and cherry-picked headline numbers; worth a major-revision round, not acceptance as-is. read the letter →

arxiv 2507.20480 v1 pith:5TT5ZVXE submitted 2025-07-28 cs.CV

Shiyang Liu , Dianyi Yang , Yu Gao , Bohan Ren , Yi Yang , Mengyin Fu This is my paper · ORCID
classification cs.CV
keywords 3DGaussianSplattingsub-mapregistrationskeletonextractionfusionellipsoid-awareconvolutionmulti-viewmappingpointcloud
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

The paper tries to establish that 3D Gaussian Splatting sub-maps built from different viewpoints, such as a drone view and a ground vehicle view, can be aligned and fused without a human choosing a reference map and without a hard distance cutoff that deletes scene geometry. Its pipeline extracts a compact skeleton from each sub-map, registers the maps with features that adapt to the anisotropic shape of each Gaussian, and then merges overlapping Gaussians with a multi-factor score. If the claim holds, multi-robot or multi-view mapping could produce a single high-fidelity 3D-GS map automatically, with the reported gains of a 41.9% reduction in rotation registration error on a complex scene and a 10.11 dB PSNR improvement on a challenging fusion scene.

What carries the argument

The load-bearing machinery is the pair formed by the Gaussian-to-Skeleton (G2D) distance and the Gaussian-Adaptive KPConv (GA-KPConv). G2D is intended to score how well each Gaussian fits a skeleton by evaluating a covariance-scaled Gaussian density at nearby skeleton points, with DBSCAN giving the initial skeleton and a curvature-regularized energy refining it. GA-KPConv selects neighborhoods by Mahalanobis distance using each Gaussian's covariance $\Sigma_i$ and morphs spherical kernel points into ellipsoidal receptive fields through $\Sigma_i^{-1}$, so local features respond to anisotropic structure; these features feed a geometric transformer that estimates the rigid transformation. The fusion step then computes, for each overlapping Gaussian, a weighted score $S_{tot} = \alpha S_{ske} + \beta S_{deta} + \gamma S_{cen}$ (with the paper's Algorithm 1 also mixing in a max condition) to retain the better primitive instead of applying a hard threshold.

What would settle it

Take two Gaussians with different means or covariances that are assigned to the same skeleton point, evaluate Eq. (1) for both, and check whether the computed values differ; if they do not, the skeleton-refinement objective is not actually per-Gaussian, so the method's alignment and fusion scores would lose their stated geometric grounding.

Watch

Extended reading notes

Core claim

The central claim is that two independently reconstructed 3D-GS sub-maps can be registered and fused automatically by exploiting the Gaussian geometry itself. A Gaussian-to-Skeleton distance, initialized by DBSCAN clustering and refined by an L1-style curvature-regularized energy, produces a structural skeleton for each sub-map; a Gaussian-adaptive KPConv uses Mahalanobis-distance neighborhoods and covariance-morphed ellipsoidal kernels to extract local features; and a soft fusion score balances skeleton adherence, local detail response, and scene-center proximity to decide which Gaussians survive in the overlap. The paper argues that this removes the manual reference-map selection and the rendering-quality loss caused by hard-threshold primitive filtering, and it reports reduced rotation and translation registration errors together with higher PSNR, SSIM, and lower LPIPS on the ScanNet-GSReg benchmark and its own Coord dataset.

Load-bearing premise

The entire skeleton pipeline assumes that Eq. (1) really measures each Gaussian's distance to the skeleton, but as printed the right-hand side sums over all Gaussians assigned to a skeleton point and never uses the index $i$, so the energy in Eq. (2) would not, taken literally, know which Gaussian is being fitted.

Editorial extensions

If this is right

  • Registration no longer requires a manually selected reference sub-map, because either sub-map's skeleton can anchor the alignment and the fusion skeleton is built by merging both.
  • Complex and anisotropic scenes should register more accurately: the paper reports a 41.9% reduction in RRE and an 82.3% reduction in RTE on scene 0089-01, at the cost of about 21% more computation time.
  • Fusion should preserve structural detail that hard-threshold filtering destroys, with a reported 10.11 dB PSNR gain and 22.3% LPIPS reduction on an occluded scene, and higher average PSNR than the GaussReg baseline.
  • Skeleton quality improves in the same pipeline: at $\epsilon=1.0$, minPts $=8$, the GA-L1 skeleton reaches 3.5% curvature deviation and 0.974 connectivity versus 7.9% and 0.920 for L1 extraction.
  • Ground-and-aerial mapping can be automated end to end, since the two sub-maps are fused into one coordinate system without manual scale or template selection.

Reading between the lines

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

  • If Eq. (1) is repaired so that the distance genuinely depends on the queried Gaussian $h_i$ (for instance, by taking the minimum over skeleton points of that Gaussian's density), the skeleton objective becomes well-defined and the reported fusion gains might be reproducible; as printed, the formula is ambiguous.
  • The same skeleton-plus-covariance machinery could plausibly serve as a loop-closure cue in incremental mapping, since a stable skeleton identity could signal when a robot re-enters a previously mapped region; the paper does not develop this use.
  • The covariance-shaped kernels suggest a natural extension to multi-sensor fusion where each primitive's covariance encodes sensor uncertainty, so LiDAR and vision maps could be registered with the same anisotropic features; this is an inference, not a claim in the paper.
  • The authors state in the conclusion that the feature extractor is geometry-only and lacks semantic awareness; adding semantic cues is a natural next step that would not require changing the skeleton or fusion machinery.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

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 presents an end-to-end framework for registering and fusing independently reconstructed 3D Gaussian Splatting (3D-GS) sub-maps. The method has three components: (i) a skeleton extraction and refinement procedure based on a proposed Gaussian-to-Skeleton distance and DBSCAN initialization; (ii) a Gaussian-adaptive KPConv feature extractor using Mahalanobis-distance neighborhoods and ellipsoid-aware kernels; and (iii) a multi-factor fusion score that balances skeleton adherence, local detail, and scene-center proximity. Experiments on the ScanNet-GSReg dataset and a self-collected Coord dataset compare the method against GaussReg and GaussReg-NF, reporting reduced rotation/translation/scale errors and improved PSNR, SSIM, and LPIPS.

Significance. If the technical content were correct, the framework would be a practically useful contribution to collaborative 3D-GS mapping: it removes manual reference-map selection and exploits Gaussian attributes beyond point positions, which are relevant for robotic perception. However, the central equations are internally inconsistent as written, the fusion algorithm does not match the described scoring scheme, and the empirical evaluation is limited to a single baseline. The paper does not provide code or machine-checked artifacts, so the reported gains cannot be independently verified. As submitted, the empirical claims are not supported.

major comments (4)
  1. [Sec. III-A, Eq. (1) and Eq. (2)] The G2D distance D(h_i,S) as written does not depend on the left-hand Gaussian h_i: the right-hand side sums over all i in G_j, where G_j is the set of Gaussians associated with skeleton point q_j, so every Gaussian in the same cluster receives the same value. Moreover, the exponential is a similarity that tends to zero as q_j moves away from the cluster means, so minimizing E(S) in Eq. (2) does not attract skeleton points toward the Gaussians; without a bounded domain, the minimum is approached at infinity. Because the skeleton refinement in Sec. III-A and the fusion scoring in Sec. III-C both inherit this quantity, the core optimization is undefined as stated.
  2. [Algorithm 1] In the fusion loop, Stot(mu) is computed as alpha*Sske(mu) + beta*Sdeta(mu) + gamma*Sske(mu), so the skeleton-adherence term appears twice and the scene-center proximity score Scen, which the text in Sec. III-C describes as one of the three factors, never enters. The final selection and the soft-thresholding condition therefore do not implement the described multi-factor fusion strategy, and the reported fusion results cannot be attributed to the method as written.
  3. [Abstract and Sec. IV-B, Table I and Table III] The headline improvements in the abstract are not representative of the average results: the 41.9% RRE reduction is the improvement on scene 0089-01 only, while the average over the displayed ScanNet-GSReg scenes is approximately 24%, and the 10.11 dB PSNR gain is from scene 0715-00 only. The authors should report aggregate statistics over the full test set, including the number of scenes and variance, and should not present single-scene gains as general improvements.
  4. [Sec. IV-B and Sec. IV-C] The registration and fusion evaluations compare the proposed method only with GaussReg-NF and GaussReg, respectively. Given the paper's claim of a general framework, a single baseline is insufficient to establish competitiveness. The authors should compare with other recent 3D-GS registration and fusion methods and provide ablations for the contributions of GA-KPConv, the skeleton term, and each fusion factor. This is especially important because the method has many free parameters (alpha, beta, gamma, tau, lambda, DBSCAN epsilon and minPts, epsilon_skel, epsilon_overlap, sigma, M) whose sensitivity is not analyzed.
minor comments (5)
  1. [Sec. III-A] The text contains a duplicated sentence: 'This weighted center weights denser Gaussians more in cluster formation...' appears twice, and the sentence about the first and second terms in Eq. (2) is repeated with identical wording, suggesting an editing error.
  2. [Sec. III-B, Eq. (4)] The subscript notation is inconsistent: the kernel weighting is written w_m, but the exponent uses d_k and x_k, and the relationship between the m and k indices is not defined.
  3. [Table III caption] The caption says 'Quantitative comparison of registration results' but the table reports rendering metrics (PSNR, SSIM, LPIPS) after fusion; the caption should say 'fusion results' instead.
  4. [Tables I and III] The Avg. rows cannot be verified from the printed rows because of the ellipsis; the authors should state the total number of test scenes and report mean and standard deviation for each metric.
  5. [Figs. 3-6] The qualitative comparisons would be easier to assess with error maps or quantitative overlays; as printed, the visual differences between methods are difficult to judge.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the reported registration and fusion gains are empirical comparisons against external baselines; the only self-citation is a non-load-bearing related-work reference.

full rationale

The central claims are experimental, not derived constants: Table I compares RRE/RTE/RSE/Time against GaussReg-NF on ScanNet-GSReg, and Table III compares PSNR/SSIM/LPIPS against GaussReg; the Coord dataset is an additional evaluation set. No equation in the paper takes a fitted parameter and renames it as a prediction; the fusion weights α, β, γ are free coefficients used in Algorithm 1, but no reported number is algebraically forced by them. The only author-overlapping citation is [4] (MC-NeRF), used in the introduction as an example of NeRF-family reconstruction and never invoked to justify the skeleton, convolution, or fusion design. The use of the extracted skeleton to guide fusion is method construction, not circular reasoning, since the fusion scores are evaluated by rendering quality rather than derived from the skeleton objective. The paper does contain two internal inconsistencies that are correctness concerns, not circularity: Eq. (1) prints D(h_i,S) with a right-hand side that sums over i∈G_j and never references the argument h_i, making the per-Gaussian distance undefined as written; and Algorithm 1 writes Stot with γS_ske twice instead of the announced γS_cen. These defects make some formulas unverifiable but do not reduce any claimed result to its own input, so the circularity score is 0.

Assumptions & free parameters 8 free parameters · 5 assumptions · 0 invented entities

The method rests on several unverified modeling choices: the suitability of L1-medial skeletons for Gaussian Splatting scenes, the reliability of GeoTransformer on GA-KPConv features, the informal curvature proxy on discrete points, the adequacy of DBSCAN initialization, and the Mahalanobis neighborhood criterion. All are reasonable starting points, but none is validated in the paper, and the G2D distance that connects them is mis-specified.

free parameters (8)
  • α (fusion weight for skeleton adherence) = not reported
    Weight in Stot selection, no ablation or value given.
  • β (fusion weight for local detail) = not reported
    Weight in Stot selection, no ablation or value given.
  • γ (fusion weight for scene-center proximity) = not reported
    Weight in Stot selection, no ablation or value given.
  • τ (detail retention threshold) = not reported
    Threshold in the final selection rule; no value given.
  • λ (curvature regularization weight) = not reported
    Weight in E(S) for skeleton smoothness; no value or sensitivity analysis.
  • DBSCAN ϵ and minPts = ϵ=0.5..1.0, minPts=4..8 (sweep in Table II); final pipeline values not reported
    Clustering parameters for skeleton initialization; only a limited sweep for skeleton metrics is shown, not for the full pipeline.
  • ε_skel and ε_overlap (skeleton merge and overlap thresholds) = not reported
    Thresholds in Algorithm 1 that define skeleton merging and overlap region; no values are given.
  • σ and M (GA-KPConv kernel radius and number of kernel points) = not reported
    Kernel parameters in Sec. III-B; no values are reported.
assumptions (5)
  • domain assumption L1-medial skeleton extraction [46] provides a meaningful structural prior for 3D-GS scenes.
    Sec. III-A builds the GA-L1 skeleton on the L1-medial skeleton framework; if the prior is unsuited to Gaussian Splatting data, the skeleton-based registration/fusion loses its foundation.
  • domain assumption GeoTransformer [22] can produce accurate relative pose estimates from GA-KPConv point features.
    Sec. IV-B uses GeoTransformer as the registration backbone; the paper provides no evidence that GeoTransformer is well-calibrated for the proposed feature space.
  • ad hoc to paper The curvature regularization term ||∇q_j||^2 approximates local curvature of a discrete point-set skeleton.
    Sec. III-A Eq. (2); for discrete skeleton points, a gradient is not well-defined, so the term is an informal proxy and is not derived or validated.
  • domain assumption DBSCAN on Gaussian centers yields a reliable initial skeleton without a predefined cluster count.
    Sec. III-A; the sensitivity of the final pipeline to DBSCAN parameters is only tested for skeleton metrics, not registration or fusion.
  • domain assumption Mahalanobis distance with per-Gaussian covariance is an appropriate neighborhood criterion for anisotropic feature extraction.
    Sec. III-B Eq. (3); this is a reasonable modeling choice but is not compared to other anisotropic neighborhood definitions.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Automated 3D-GS Registration and Fusion via Skeleton Alignment and Gaussian-Adaptive Features." pith.science (2026). https://pith.science/paper/5TT5ZVXE

@misc{pith2026250720480,
  author       = {Pith},
  title        = {Pith review of: Automated 3D-GS Registration and Fusion via Skeleton Alignment and Gaussian-Adaptive Features},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/5TT5ZVXE}},
  note         = {Machine review of arXiv:2507.20480}
}
read the original abstract

In recent years, 3D Gaussian Splatting (3D-GS)-based scene representation demonstrates significant potential in real-time rendering and training efficiency. However, most existing methods primarily focus on single-map reconstruction, while the registration and fusion of multiple 3D-GS sub-maps remain underexplored. Existing methods typically rely on manual intervention to select a reference sub-map as a template and use point cloud matching for registration. Moreover, hard-threshold filtering of 3D-GS primitives often degrades rendering quality after fusion. In this paper, we present a novel approach for automated 3D-GS sub-map alignment and fusion, eliminating the need for manual intervention while enhancing registration accuracy and fusion quality. First, we extract geometric skeletons across multiple scenes and leverage ellipsoid-aware convolution to capture 3D-GS attributes, facilitating robust scene registration. Second, we introduce a multi-factor Gaussian fusion strategy to mitigate the scene element loss caused by rigid thresholding. Experiments on the ScanNet-GSReg and our Coord datasets demonstrate the effectiveness of the proposed method in registration and fusion. For registration, it achieves a 41.9\% reduction in RRE on complex scenes, ensuring more precise pose estimation. For fusion, it improves PSNR by 10.11 dB, highlighting superior structural preservation. These results confirm its ability to enhance scene alignment and reconstruction fidelity, ensuring more consistent and accurate 3D scene representation for robotic perception and autonomous navigation.

Figures

Figures reproduced from arXiv: 2507.20480 by the authors.

Figure 1
Figure 1. Illustration of our task. A vehicle (Diablo) and a drone (Tello) independently capture data and reconstruct separate 3D-GS sub-map. The drone view offers broad structural coverage, while the vehicle view captures fine details. A fusion strategy integrates both, producing a unified 3D-GS model with global completeness and high fidelity. and fusion of multiple sub-maps. As illustrated in Fig.1, when an unmanned system… view at source ↗
Figure 2
Figure 2. Overview of the proposed method. Given two sub-maps collected from different perspectives, we extract skeletons and local details [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Qualitative comparison of registration results on the ScanNet-GSReg dataset. [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Visualization of feature extraction results on 3D-GS scenes. Warmer colors (red) indicate richer local details [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]
Figure 5
Figure 5. Figure 5: Iterative optimization of the GA-L1 skeleton. The first row depicts the stepwise refinement of the drone view skeleton, while the second row shows the optimization process of the vehicle view skeleton [PITH_FULL_IMAGE:figures/full_fig_p006_5.png]
Figure 6
Figure 6. Figure 6: Qualitative comparison of fusion results on the ScanNet [PITH_FULL_IMAGE:figures/full_fig_p007_6.png]

Review history (2 revisions) →

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

47 extracted references · 40 canonical work pages

  1. [1]

    T.-Y . Lin, S. Belongie, and J. Hays. Cross-view image geolocalization. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pages 891–898, 2013

    work page 2013

  2. [2]

    Mildenhall, P

    B. Mildenhall, P. P. Srinivasan, M. Tancik, J. T. Barron, R. Ramamoor- thi, and R. Ng. Nerf: Representing scenes as neural radiance fields for view synthesis. Communications of the ACM , 65(1):99–106, 2021

    work page 2021

  3. [3]

    J. T. Barron, B. Mildenhall, M. Tancik, P. Hedman, R. Martin-Brualla, and P. P. Srinivasan. Mip-nerf: A multiscale representation for anti- aliasing neural radiance fields. In Proceedings of the IEEE/CVF International Conference on Computer Vision , pages 5855–5864, 2021

    work page 2021

  4. [4]

    Y . Gao, L. Su, H. Liang, Y . Yue, Y . Yang, and M. Fu. Mc-nerf: Multi- camera neural radiance fields for multi-camera image acquisition systems. arXiv preprint arXiv:2309.07846 , 2023

    work page Pith review arXiv 2023

  5. [5]

    Kerbl, G

    B. Kerbl, G. Kopanas, Thomas Leimk ¨uhler, and G. Drettakis. 3d gaussian splatting for real-time radiance field rendering. ACM Trans. Graph., 42(4):139–1, 2023

    work page 2023

  6. [6]

    G. Wu, T. Yi, J. Fang, L. Xie, X. Zhang, W. Wei, W. Liu, Q. Tian, and X. Wang. 4d gaussian splatting for real-time dynamic scene rendering. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 20310–20320, 2024

    work page 2024

  7. [7]

    M. Qin, W. Li, J. Zhou, H. Wang, and H. Pfister. Langsplat: 3d lan- guage gaussian splatting. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages 20051–20060, 2024

    work page 2024

  8. [8]

    Zheng, X

    Y . Zheng, X. Chen, Y . Zheng, S. Gu, R. Yang, B. Jin, P. Li, C. Zhong, Z. Wang, L. Liu, et al. Gaussiangrasper: 3d language gaussian splatting for open-vocabulary robotic grasping. IEEE Robotics and Automation Letters, 2024

    work page 2024

Show all 47 references
  1. [9]

    J. Wang, Q. Zhang, J. Sun, J. Cao, G. Han, W. Zhao, W. Zhang, Y . Shao, Y . Guo, and R. Xu. Reinforcement learning with generalizable gaussian splatting. In 2024 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) , pages 435–441. IEEE, 2024

    2024

  2. [10]

    Liv-gaussmap: Lidar-inertial-visual fusion for real-time 3d radiance field map rendering

    Sheng Hong, Junjie He, Xinhu Zheng, Chunran Zheng, and Shaojie Shen. Liv-gaussmap: Lidar-inertial-visual fusion for real-time 3d radiance field map rendering. IEEE Robotics and Automation Letters , 2024

    2024

  3. [11]

    Matsuki, R

    H. Matsuki, R. Murai, P. H. J. Kelly, and A. J. Davison. Gaussian splatting slam. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages 18039–18048, 2024

    2024

  4. [12]

    C. Yan, D. Qu, D. Xu, B. Zhao, Z. Wang, D. Wang, and X. Li. Gs- slam: Dense visual slam with 3d gaussian splatting. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recog- nition, pages 19595–19604, 2024

    2024

  5. [13]

    Chang, Y

    J. Chang, Y . Xu, Y . Li, Y . Chen, W. Feng, and X. Han. Gaussreg: Fast 3d registration with gaussian splatting. In European Conference on Computer Vision , pages 407–423. Springer, 2024

    2024

  6. [14]

    Z. Yuan, T. Zhang, M. Johnson-Roberson, and W. Zhi. Photoreg: Photometrically registering 3d gaussian splatting models. arXiv preprint arXiv:2410.05044, 2024

    2024 arXiv

  7. [15]

    L. Zhu, Y . Li, E. Sandstr ¨om, S. Huang, K. Schindler, and I. Armeni. Loopsplat: Loop closure by registering 3d gaussian splats. arXiv preprint arXiv:2408.10154, 2024

    2024 arXiv

  8. [16]

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

    2020

  9. [17]

    H. Li, A. Liu, X. Xie, H. Guo, H. Xiong, and X. Zheng. Learning dense consistent features for aerial-to-ground structure-from-motion. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing , 16:5089–5102, 2023

    2023

  10. [18]

    Z. Hou, Y . Liu, and L. Zhang. Pos-gift: A geometric and intensity- invariant feature transformation for multimodal images. Information Fusion, 102:102027, 2024

    2024

  11. [19]

    J. Liu, H. Yin, B. Liu, and P. Lu. Tie point matching between terrestrial and aerial images based on patch variational refinement. Remote Sensing, 15(4):968, 2023

    2023

  12. [20]

    J. Li, Q. Hu, and M. Ai. Point cloud registration based on one-point ransac and scale-annealing biweight estimation. IEEE Transactions on Geoscience and Remote Sensing , 59(11):9716–9729, 2021

    2021

  13. [21]

    Y . Yuan, Y . Wu, X. Fan, M. Gong, W. Ma, and Q. Miao. Egst: Enhanced geometric structure transformer for point cloud registra- tion. IEEE Transactions on Visualization and Computer Graphics , 30(9):6222–6234, 2023

    2023

  14. [22]

    Z. Qin, H. Yu, C. Wang, Y . Guo, Y . Peng, S. Ilic, D. Hu, and K. Xu. Geotransformer: Fast and robust point cloud registration with geometric transformer. IEEE Transactions on Pattern Analysis and Machine Intelligence, 45(8):9806–9821, 2023

    2023

  15. [23]

    Poiesi and D

    F. Poiesi and D. Boscaini. Learning general and distinctive 3d local deep descriptors for point cloud registration. IEEE Transactions on Pattern Analysis and Machine Intelligence , 45(3):3979–3985, 2022

    2022

  16. [24]

    J. Yu, L. Ren, Y . Zhang, W. Zhou, L. Lin, and G. Dai. Peal: Prior-embedded explicit attention learning for low-overlap point cloud registration. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages 17702–17711, 2023

    2023

  17. [25]

    Shorinwa, J

    O. Shorinwa, J. Sun, M. Schwager, and A. Majumdar. Siren: Semantic, initialization-free registration of multi-robot gaussian splatting maps. arXiv preprint arXiv:2502.06519 , 2025

    2025 arXiv

  18. [26]

    On the generalized distance in statistics

    Prasanta Chandra Mahalanobis. On the generalized distance in statistics. Sankhy¯a: The Indian Journal of Statistics, Series A (2008-) , 80:S1–S7, 2018

    2008

  19. [27]

    Q. Lai, Q. Xin, Y . Tian, X. Chen, Y . Li, and R. Wu. Structural analysis and 3d reconstruction of underground pipeline systems based on lidar point clouds. Remote Sensing , 17(2):341, 2025

    2025

  20. [28]

    Rodrigues, M

    M. Rodrigues, M. Kormann, C. Schuhler, and P. Tomek. Structured light techniques for 3d surface reconstruction in robotic tasks. In Pro- ceedings of the 8th International Conference on Computer Recognition Systems CORES 2013 , pages 805–814. Springer, 2013

    2013

  21. [29]

    P. Li, R. Wang, Y . Wang, and W. Tao. Evaluation of the icp algorithm in 3d point cloud registration. IEEE Access , 8:68030–68048, 2020

    2020

  22. [30]

    H. Liu, T. Liu, Y . Li, M. Xi, T. Li, and Y . Wang. Point cloud registration based on mcmc-sa icp algorithm. IEEE Access , 7:73637– 73648, 2019

    2019

  23. [31]

    Vizzo, T

    I. Vizzo, T. Guadagnino, B. Mersch, L. Wiesmann, J. Behley, and C. Stachniss. Kiss-icp: In defense of point-to-point icp–simple, accurate, and robust registration if done the right way. IEEE Robotics and Automation Letters , 8(2):1029–1036, 2023

    2023

  24. [32]

    Cheng, Z

    Y . Cheng, Z. Huang, S. Quan, X. Cao, S. Zhang, and J. Yang. Sampling locally, hypothesis globally: Accurate 3d point cloud registration with a ransac variant. Visual Intelligence, 1(1):20, 2023

    2023

  25. [33]

    Chung and W.-T

    K.-L. Chung and W.-T. Chang. Centralized ransac based point cloud registration with fast convergence and high accuracy. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing , 2024

    2024

  26. [34]

    W. Lu, G. Wan, Y . Zhou, X. Fu, P. Yuan, and S. Song. Deepvcp: An end-to-end deep neural network for point cloud registration. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 12–21, 2019

    2019

  27. [35]

    F. Lu, G. Chen, Y . Liu, L. Zhang, S. Qu, S. Liu, and R. Gu. Hregnet: A hierarchical network for large-scale outdoor lidar point cloud registration. In Proceedings of the IEEE/CVF International Conference on Computer Vision , pages 16014–16023, 2021

    2021

  28. [36]

    Y . Wu, P. Gong, Y . Yuan, M. Gong, W. Ma, and Q. Miao. Evolutionary neural architecture search framework with masked encoding mecha- nism for point cloud registration. IEEE Transactions on Emerging Topics in Computational Intelligence , 2025

    2025

  29. [37]

    L. Goli, D. Rebain, S. Sabour, A. Garg, and A. Tagliasacchi. Nerf2nerf: Pairwise registration of neural radiance fields. In 2023 IEEE International Conference on Robotics and Automation (ICRA) , pages 9354–9361. IEEE, 2023

    2023

  30. [38]

    Chen and G

    Y . Chen and G. H. Lee. Dreg-nerf: Deep registration for neural radi- ance fields. In Proceedings of the IEEE/CVF International Conference on Computer Vision , pages 22703–22713, 2023

    2023

  31. [39]

    J. Fang, S. Lin, I. Vasiljevic, V . Guizilini, R. Ambrus, A. Gaidon, G. Shakhnarovich, and M. R. Walter. Nerfuser: Large-scale scene representation by nerf fusion. arXiv preprint arXiv:2305.13307 , 2023

    2023 arXiv

  32. [40]

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

    2019

  33. [41]

    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 , page 123, 2020

    2020

  34. [42]

    Zhang, B

    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

    2024 arXiv

  35. [43]

    Y . Ham, M. Michalkiewicz, and G. Balakrishnan. Dragon: Drone and ground gaussian splatting for 3d building reconstruction. In 2024 IEEE International Conference on Computational Photography (ICCP), pages 1–12. IEEE, 2024

    2024

  36. [44]

    G. Kang, J. Yoo, J. Park, S. Nam, H. Im, S. Kim, E. Park, et al. Self- splat: Pose-free and 3d prior-free generalizable 3d gaussian splatting. arXiv preprint arXiv:2411.17190 , 2024

    2024 arXiv

  37. [45]

    Zhang, Y

    C. Zhang, Y . Cao, and L. Zhang. Crossview-gs: Cross-view gaus- sian splatting for large-scale scene reconstruction. arXiv preprint arXiv:2501.01695, 2025

    2025 arXiv

  38. [46]

    Huang, S

    H. Huang, S. Wu, D. Cohen-Or, M. Gong, H. Zhang, G. Li, and B. Chen. L1-medial skeleton of point cloud. ACM Trans. Graph. , 32(4):65–1, 2013

    2013

  39. [47]

    Ester, H.-P

    M. Ester, H.-P. Kriegel, J. Sander, X. Xu, et al. A density-based algorithm for discovering clusters in large spatial databases with noise. In KDD, volume 96, pages 226–231, 1996

    1996

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.