Pith. sign in

REVIEW 5 major objections 6 minor 91 references

Hyperbolic Chamfer Distance for Point Cloud Completion and Beyond

T0 review · 5 major / 6 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read Replacing Chamfer Distance with a hyperbolic arcosh loss improves point cloud completion across benchmarks.

desk verdict A one-line loss tweak with consistent but unproven gains: the hyperbolic framing is mostly decoration, and the scale confound is real, but it's worth refereeing. read the letter →

arxiv 2412.17951 v1 pith:ZPQMQNCY submitted 2024-12-23 cs.CV

Fangzhou Lin , Songlin Hou , Haotian Liu , Shang Gao , Kazunori D Yamada , Haichong K. Zhang , Ziming Zhang This is my paper · ORCID
classification cs.CV
keywords pointcloudcompletionChamferdistancehyperbolicspacearcoshlossdensity-awareupsamplingsingle-viewreconstruction
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 proposes Hyperbolic Chamfer Distance (HyperCD), a drop-in replacement for the Chamfer Distance loss in point cloud completion. HyperCD computes the usual nearest-neighbor Chamfer sum but passes each squared Euclidean distance through $\operatorname{arcosh}(1+\alpha r^2)$, with $\alpha>0$. The claim is that this simple change reweights gradients so that already-close point pairs are preserved while poor matches are corrected gradually, and that models trained with it outperform the same models trained with CD or DCD. On the PCN, ShapeNet-55, and ShapeNet-34 benchmarks the paper reports state-of-the-art completion results, plus gains in single-image reconstruction and point cloud upsampling. The authors argue this is the first use of hyperbolic geometry in point cloud completion.

What carries the argument

The central object is the HyperCD distance $d(x_{ij},y_{ik})=\operatorname{arcosh}(1+\alpha\|x_{ij}-y_{ik}\|^2)$, applied inside the Chamfer min-sum over nearest neighbors. Its load-bearing property is the gradient weight $z_{ij}=2\alpha r/\sqrt{(1+\alpha r^2)^2-1}$, which Proposition 4 shows is strictly decreasing in the Euclidean match distance $r$; this is what redirects optimization toward preserving good matches. Proposition 1 justifies computing the minimum in Euclidean space first, keeping the algorithm almost as cheap as ordinary Chamfer Distance.

What would settle it

Train the same completion backbone with the squared-Euclidean loss $d(r)=r^2$ and with HyperCD $\operatorname{arcosh}(1+\alpha r^2)$ while matching the initial gradient magnitude; if the squared-CD model matches HyperCD's benchmark error, then the arcosh nonlinearity and its specific weighting curve are not the operative cause. A second check is to sweep $\alpha$ widely on a held-out set to see whether the reported results depend on a narrow tuned range.

Watch

Extended reading notes

Core claim

HyperCD replaces the per-pair distance $d(x_{ij}, y_{ik})$ inside Chamfer Distance with $\operatorname{arcosh}(1+\alpha\|x_{ij}-y_{ik}\|^2)$, $\alpha>0$. Because $\operatorname{arcosh}(1+\alpha r^2)$ is strictly increasing in $r$, the nearest-neighbor match found in Euclidean space is unchanged, so the extra cost is just one $\operatorname{arcosh}$ per match. During backpropagation the gradient carries a factor $z_{ij}=2\alpha r/\sqrt{(1+\alpha r^2)^2-1}$ that decreases as the matched distance $r$ grows; the paper's Proposition 4 proves this monotone decrease. The intended effect is that accurate matches receive larger gradient weight and are stabilized, while distant outlier matches receive smaller weight and are adjusted incrementally. Empirically the paper shows this weighting yields lower Chamfer error and smoother surfaces than CD and DCD across several backbones, and reports state-of-the-art numbers on PCN, ShapeNet-55, and ShapeNet-34.

Load-bearing premise

The load-bearing premise is that HyperCD's distance-dependent gradient weighting, not incidental differences like loss scale or the grid-searched $\alpha$, causes the reported improvements; the paper does not compare against other monotone transforms or rescaled CD baselines to rule those out.

Editorial extensions

If this is right

  • Training any of the tested completion backbones (PCN, FoldingNet, PMP-Net, PoinTr, SnowflakeNet, PointAttN, SeedFormer) with HyperCD in place of CD lowers per-point Chamfer error on PCN.
  • On ShapeNet-55 and ShapeNet-34, HyperCD with SeedFormer improves both CD and F-Score@1% over the CD-trained baseline, including on unseen categories.
  • HyperCD helps downstream generative tasks: swapping CD for HyperCD in SnowflakeNet improves single-view reconstruction, and swapping it into four upsampling networks improves CD, Hausdorff, and point-to-surface metrics on PU1K.
  • HyperCD's computational cost is close to that of CD: the only added operation per match is one arcosh, in contrast to the heavier full hyperbolic distance.

Reading between the lines

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

  • Because the paper compares HyperCD against CD and DCD but not against other monotone transforms (e.g., squared Euclidean CD or a logarithmic loss) with matched gradient scales, the specific arcosh form may not be essential; a reweighting family with similar close-pair emphasis could reproduce the gains.
  • A testable extension is to use the same position-aware weighting as a plug-in for other set-matching losses (e.g., in 2D/3D registration or shape retrieval), where the same outlier-versus-fine-detail trade-off occurs.
  • The reported dependence of optimal $\alpha$ on learning rate suggests that loss-scale calibration, rather than geometry, may drive part of the improvement; ablating with per-backbone learning-rate rescaling would separate the two effects.
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

5 major / 6 minor

Summary. The paper proposes Hyperbolic Chamfer Distance (HyperCD), defined as arcosh(1 + α||x − y||^2) applied to each nearest-neighbor pair in the standard Chamfer Distance loss. The authors argue that this monotone reweighting gives larger gradients to close, well-matched point pairs and smaller gradients to distant, poorly matched pairs, which they call a position-aware mechanism. They report consistent improvements when HyperCD replaces CD in several point cloud completion networks on PCN, ShapeNet-55, and ShapeNet-34, and they also apply HyperCD to single-image reconstruction and point cloud upsampling. The paper includes propositions about the interchange of min and g, the monotonicity of the distance, and the limiting behavior of the gradient, plus an efficient algorithm whose cost is comparable to CD.

Significance. If the central claim is sustained, HyperCD is an attractively simple drop-in replacement for the Chamfer Distance loss: it requires one additional arcosh operation, preserves the nearest-neighbor matching step of CD, and is accompanied by public code and a small set of analytical propositions. The empirical gains are modest but appear consistent across multiple backbones and tasks, which makes the paper potentially useful to practitioners. However, the causal explanation offered for the gains is not established: Proposition 1 shows that the nearest-neighbor correspondences are identical to those of Euclidean CD, so the hyperbolic-space matching correction illustrated in Fig. 1 is not realized, and the reported experiments do not control for the extra loss-scale/learning-rate knob introduced by grid-searching α. Some reported baselines (notably DCD and KITTI Fidelity) are implausible, which weakens confidence in the loss-comparison setup.

major comments (5)
  1. [Table 4, Sec. 4.1] The KITTI results contain impossible Fidelity values: the CD-ℓ2 Fidelity row reports 0.000 for both PoinTr and HyperCD + PoinTr, and hyperCD+FoldingNet drops from 7.467 to 2.214. A Fidelity of exactly 0.000 for a learned completion model on real LiDAR scans is not credible as a per-point accuracy measure and suggests a broken evaluation pipeline or a typo in the table. Please correct or explain these numbers; as printed, they vitiate the claim that HyperCD consistently improves performance on KITTI.
  2. [Table 1, Table 7] The DCD comparisons are implausible and undermine the loss-comparison setup. In Table 1, DCD + SeedFormer reports 24.52 average L1-CD versus 6.74 for CD + SeedFormer, a factor-of-3.6 degradation that is inconsistent with the published DCD results and with the moderate numbers reported elsewhere. In Table 7, DCD gives 5.74 on ShapeNet-Part while L1-CD gives 4.16, again contradicting the original DCD paper's claims. Please verify the DCD implementation and report the original authors' numbers or explain the discrepancy.
  3. [Eq. (7), Sec. 4 Implementation, Fig. 5] The central causal claim that HyperCD's gradient weighting improves completion is not separated from the effect of loss scale. For small matched-pair distances, which are the typical regime in completion, Eq. (7) gives z ≈ √(2α), so HyperCD behaves like √(2α) times the L1-CD gradient. The paper grid-searches α while explicitly keeping learning rates and all other hyperparameters at baseline values, so α serves as an extra effective learning-rate/scale knob that the CD baselines do not have; Fig. 5 confirms that α and lr jointly affect CD. To support the position-aware explanation, the authors should compare against a rescaled L1-CD (e.g., L1-CD with a tuned multiplicative scale) and against a CD trained with a learning rate tuned in an equally favorable manner.
  4. [Proposition 1, Eq. (8), Fig. 1] Proposition 1 shows that because arcosh(1+αr²) is strictly increasing, the nearest-neighbor matches under HyperCD are exactly the Euclidean nearest-neighbor matches. Consequently, HyperCD never changes the correspondences, contradicting the Fig. 1 motivation that 'the mismatched point pairs in Euclidean space may be corrected' in hyperbolic space. The only difference from CD is the per-pair gradient weighting in Eq. (7). The paper should either reframe the contribution as a gradient-weighting scheme (dropping the matching-correction narrative) or provide evidence that the weighting mechanism, not scale, is what drives the gains, e.g., by comparing against another monotone transform with the same small-distance slope but a different large-distance profile.
  5. [Sec. 4, Tables 1-3] No error bars, confidence intervals, or multiple-seed results are reported. Several of the claimed improvements are small (e.g., Table 1: SeedFormer from 6.74 to 6.54; Table 2: SeedFormer CD-Avg from 0.92 to 0.86), and without run-to-run variance it is impossible to judge whether these differences are statistically meaningful. Please report mean and standard deviation over at least three seeds for the main completion tables, or otherwise justify the stability of the single-run numbers.
minor comments (6)
  1. [Abstract/Introduction] The phrase 'we also provide the provide experimental results beyond completion task' appears in both the abstract and the introduction; please remove the duplicated 'provide'.
  2. [Sec. 1 Contributions] The third bullet contains the typo 'HypperCD'; it should read 'HyperCD'.
  3. [Sec. 4, Upsampling] The text says 'point cloud unsampling' in the evaluation-metrics paragraph; this should be 'upsampling'.
  4. [Table 5 caption] The caption states that PU-GCN's superior performance is 'denoted as bold in the results', but no entries in the table are bold.
  5. [Sec. 4.2 Hyperparameters] The paper performs a grid search over α but never reports the chosen α values for any experiment; please report the selected values for reproducibility.
  6. [Sec. 4.2 Computation] The reported per-iteration timings (0.4239±0.0019, 0.4298±0.0014, 0.5335±0.0368 seconds) would be more interpretable if the hardware and batch size were stated in the same paragraph.

Circularity Check

0 steps flagged · score 2.0 of 10

No significant circularity: HyperCD's gradient property is definitional but the performance claims rest on external benchmarks; the main weaknesses are experimental and interpretational, not circular.

full rationale

The paper's derivation chain is self-contained. HyperCD is explicitly defined in Eq. 5 as d(xij,yik)=arcosh(1+alpha*||xij-yik||^2), and the gradient weight in Eq. 7 is computed directly from that definition. The decreasing-weight property is a mathematical consequence of the chosen beta=2 form, not a fitted quantity that is later relabeled as a prediction. The claimed state-of-the-art results on PCN, ShapeNet-55/34, KITTI, SVR, and PU1K are external benchmark evaluations, so the empirical claim is independently testable rather than forced by construction. Proposition 1 shows that nearest-neighbor matches under HyperCD coincide with Euclidean matches, which undermines the Fig. 1 motivation about hyperbolic space correcting mismatched pairs; this is an internal-consistency problem, not a circular derivation. Likewise, the causal attribution to the weighting mechanism is not tested against a rescaled L1-CD baseline, and the grid search over alpha on the same benchmarks without reporting chosen alpha values or error bars is a selection-bias and experimental-control concern. The main self-citations (CP-Net [82] as an analysis backbone and [71] for gyrovector background) are not load-bearing for the central claim. Overall, no derivation step reduces to its own inputs, so the circularity score is low.

Assumptions & free parameters 1 free parameters · 3 assumptions · 0 invented entities

The central contribution is a new loss with one tuned hyperparameter (alpha) and an ad hoc choice of arcosh as the nonlinearity. The claimed hyperbolic-space connection is largely rhetorical because constant alpha is not a hyperbolic distance. No new entities or forces are introduced.

free parameters (1)
  • alpha = not reported (grid-searched per network and dataset)
    Hyperparameter in the HyperCD definition (Eq. 5), controlling the scale of the arcosh transform. The paper states that the tuning engineer's role is a grid search over alpha, but the selected values are not listed, so the reported results are conditional on these undisclosed settings.
assumptions (3)
  • ad hoc to paper The constant-alpha arcosh loss (Eq. 5) preserves the benefits of hyperbolic space.
    Eq. 5 is presented as a special case of the hyperbolic distance (Eq. 3) with alpha depending on point norms, but a constant alpha is not a hyperbolic distance. The paper assumes this simplification retains the motivation without testing it.
  • ad hoc to paper The gradient-weighting profile of arcosh (large weights for close pairs) is the cause of improved completion.
    This causal link is asserted after observing performance improvements; no controlled comparison against other monotone transforms or rescaled CD is performed.
  • standard math Standard backpropagation and existing network implementations are trusted.
    The paper relies on PyTorch and common training practices without new theoretical guarantees.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Hyperbolic Chamfer Distance for Point Cloud Completion and Beyond." pith.science (2026). https://pith.science/paper/ZPQMQNCY

@misc{pith2026241217951,
  author       = {Pith},
  title        = {Pith review of: Hyperbolic Chamfer Distance for Point Cloud Completion and Beyond},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ZPQMQNCY}},
  note         = {Machine review of arXiv:2412.17951}
}
read the original abstract

Chamfer Distance (CD) is widely used as a metric to quantify difference between two point clouds. In point cloud completion, Chamfer Distance (CD) is typically used as a loss function in deep learning frameworks. However, it is generally acknowledged within the field that Chamfer Distance (CD) is vulnerable to the presence of outliers, which can consequently lead to the convergence on suboptimal models. In divergence from the existing literature, which largely concentrates on resolving such concerns in the realm of Euclidean space, we put forth a notably uncomplicated yet potent metric specifically designed for point cloud completion tasks: {Hyperbolic Chamfer Distance (HyperCD)}. This metric conducts Chamfer Distance computations within the parameters of hyperbolic space. During the backpropagation process, HyperCD systematically allocates greater weight to matched point pairs exhibiting reduced Euclidean distances. This mechanism facilitates the preservation of accurate point pair matches while permitting the incremental adjustment of suboptimal matches, thereby contributing to enhanced point cloud completion outcomes. Moreover, measure the shape dissimilarity is not solely work for point cloud completion task, we further explore its applications in other generative related tasks, including single image reconstruction from point cloud, and upsampling. We demonstrate state-of-the-art performance on the point cloud completion benchmark datasets, PCN, ShapeNet-55, and ShapeNet-34, and show from visualization that HyperCD can significantly improve the surface smoothness, we also provide the provide experimental results beyond completion task.

Figures

Figures reproduced from arXiv: 2412.17951 by the authors.

Figure 1
Figure 1. Illustration of point matching in the (a) Euclidean space and (b) hyperbolic space. With the position-aware embeddings in hyperbolic space, the mismatched point pairs in Euclidean space may be corrected, leading to better completion performance. unsupervised learning[13], [14], [15], [16], [17], [18]. Super￾vised learning, utilizing a general encoder-decoder archi￾tecture, has emerged as the predominant paradigm amo… view at source ↗
Figure 2
Figure 2. Illustration of (left) some distance metrics and (right) their corresponding gradients, where the dotted curves are used in ℓ1 and ℓ2-CD, the dash ones are used in density-aware Chamfer distance (DCD) [25], and the solid curves are special cases of our HyperCD. The superiority of attention mechanisms, such as the Transformer [41], lies in their capability to effectively capture long-range interactions, unlike the li… view at source ↗
Figure 3
Figure 3. Illustration of the gradient weights using our HyperCD. All the numbers [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Visual comparison of point cloud completion results on PCN. [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]
Figure 5
Figure 5. Figure 5: The L1-CD with different α and lr. distance. Numerically, it takes 0.4239±0.0019, 0.4298±0.0014 and 0.5335±0.0368 second per iteration for training CP-Net with CD, HyperCD and hyperbolic distance, respectively. Point Correspondences. We also provide an visualization on…
Figure 6
Figure 6. Figure 6: Illustration of point correspondence change over epochs. [PITH_FULL_IMAGE:figures/full_fig_p010_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

91 extracted references · 69 canonical work pages

  1. [1]

    Pointattn: You only need attention for point cloud completion,

    J. Wang, Y. Cui, D. Guo, J. Li, Q. Liu, and C. Shen, “Pointattn: You only need attention for point cloud completion,”arXiv preprint arXiv:2203.08485, 2022. 1, 3, 6, 7

    arXiv 2022

  2. [2]

    Completing partial point clouds with outliers by collaborative completion and segmentation,

    C. Ma, Y. Yang, J. Guo, C. Wang, and Y. Guo, “Completing partial point clouds with outliers by collaborative completion and segmentation,” arXiv preprint arXiv:2203.09772, 2022. 1

    arXiv 2022

  3. [3]

    Temporal point cloud completion with pose disturbance,

    J. Shi, L. Xu, P. Li, X. Chen, and S. Shen, “Temporal point cloud completion with pose disturbance,”IEEE Robotics and Automation Letters, vol. 7, no. 2, pp. 4165–4172, 2022. 1

    2022

  4. [4]

    Epar: An efficient and privacy- aware augmented reality framework for indoor location-based services,

    Z. Peng, S. Hou, and Y. Yuan, “Epar: An efficient and privacy- aware augmented reality framework for indoor location-based services,” in 2022 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE, 2022, pp. 8948–8955. 1

    2022

  5. [5]

    Trunk volume prediction of individual populus euphratica trees based on point clouds analysis,

    Y. Huang, S. Hou, H. Ling, and H. Xu, “Trunk volume prediction of individual populus euphratica trees based on point clouds analysis,”Ecological Indicators, vol. 95, pp. 964–971, 2018. 1

    2018

  6. [6]

    Low-costpersonalized chocolate 3d printing platform,

    J.Tong,L.Zhang,S.Hou,Y.Fan,andT.Li,“Low-costpersonalized chocolate 3d printing platform,”Journal of Computer-Aided Design & Computer Graphics, vol. 27, no. 6, pp. 984–991, 2015. 1

    2015

  7. [7]

    Pu-net: Point cloud upsampling network,

    L. Yu, X. Li, C.-W. Fu, D. Cohen-Or, and P.-A. Heng, “Pu-net: Point cloud upsampling network,” inProceedings of the IEEE conference on computer vision and pattern recognition, 2018, pp. 2790–2799. 1, 3, 7, 9

    2018

  8. [8]

    Point cloud upsampling via disentangled refinement,

    R. Li, X. Li, P.-A. Heng, and C.-W. Fu, “Point cloud upsampling via disentangled refinement,” inProceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2021, pp. 344–353. 1

    2021

Show all 91 references
  1. [9]

    Score-based point cloud denoising,

    S. Luo and W. Hu, “Score-based point cloud denoising,” inPro- ceedingsoftheIEEE/CVFInternationalConferenceonComputerVision , 2021, pp. 4583–4592. 1

    2021

  2. [10]

    High-fidelity point cloud completion with low-resolution recovery and noise- aware upsampling,

    R.-W. Li, B. Wang, C.-P. Li, L.-X. Zhang, and L. Gao, “High-fidelity point cloud completion with low-resolution recovery and noise- aware upsampling,”arXiv preprint arXiv:2112.11271, 2021. 1

    2021 arXiv

  3. [11]

    Seedformer: Patch seeds based point cloud completion with upsample transformer,

    H. Zhou, Y. Cao, W. Chu, J. Zhu, T. Lu, Y. Tai, and C. Wang, “Seedformer: Patch seeds based point cloud completion with upsample transformer,”arXiv preprint arXiv:2207.10315, 2022. 1, 6, 7, 8

    2022 arXiv

  4. [12]

    Denoise and contrast for category agnostic shape completion,

    A. Alliegro, D. Valsesia, G. Fracastoro, E. Magli, and T. Tommasi, “Denoise and contrast for category agnostic shape completion,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2021, pp. 4629–4638. 1

    2021

  5. [13]

    Pcn: point completion network,

    W. Yuan, T. Khot, D. Held, C. Mertz, and M. Hebert, “Pcn: point completion network,” in3DV, 2018. 1, 2, 6, 8

    2018

  6. [14]

    Cascaded refinement network for point cloud completion,

    X. Wang, M. H. A. J. , and G. H. Lee, “Cascaded refinement network for point cloud completion,” inCVPR, 2020. 1

    2020

  7. [15]

    Self-supervisedpoint cloud completion via inpainting,

    H.Mittal,B.Okorn,A.Jangid,andD.Held,“Self-supervisedpoint cloud completion via inpainting,”arXiv preprint arXiv:2111.10701,

    arXiv

  8. [16]

    Learning a structured latent space for unsupervised point cloud comple- tion,

    Y.Cai,K.-Y.Lin,C.Zhang,Q.Wang,X.Wang,andH.Li,“Learning a structured latent space for unsupervised point cloud comple- tion,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2022, pp. 5543–5553. 1

    2022

  9. [17]

    Reconstruction- aware prior distillation for semi-supervised point cloud comple- tion,

    Z. Fan, Y. He, Z. Wang, K. Wu, H. Liu, and J. He, “Reconstruction- aware prior distillation for semi-supervised point cloud comple- tion,” arXiv preprint arXiv:2204.09186, 2022. 1

    2022 arXiv

  10. [18]

    Self-supervised point cloud completion on real traffic scenes via scene-concerned bottom-up mechanism,

    Y. Ren, P. Cong, X. Zhu, and Y. Ma, “Self-supervised point cloud completion on real traffic scenes via scene-concerned bottom-up mechanism,” in 2022 IEEE International Conference on Multimedia and Expo (ICME). IEEE, 2022, pp. 1–6. 1

    2022

  11. [19]

    Pointr: Diverse point cloud completion with geometry-aware transformers,

    X. Yu, Y. Rao, Z. Wang, Z. Liu, J. Lu, and J. Zhou, “Pointr: Diverse point cloud completion with geometry-aware transformers,” in ICCV, 2021. 1, 3, 6, 8

    2021

  12. [20]

    Snowflakenet:Pointcloudcompletionbysnowflakepoint deconvolution with skip-transformer,

    P. Xiang, X. Wen, Y.-S. Liu, Y.-P. Cao, P. Wan, W. Zheng, and Z.Han,“Snowflakenet:Pointcloudcompletionbysnowflakepoint deconvolution with skip-transformer,” inICCV, 2021. 1, 3, 6, 7, 10

    2021

  13. [21]

    Comprehensive review of deep learning-based 3d point cloud completion processing and analysis,

    B. Fei, W. Yang, W.-M. Chen, Z. Li, Y. Li, T. Ma, X. Hu, and L. Ma, “Comprehensive review of deep learning-based 3d point cloud completion processing and analysis,”IEEE Transactions on Intelligent Transportation Systems, 2022. 1

    2022

  14. [22]

    Fpcc: Fast point JOURNAL OF LATEX CLASS FILES, VOL. 14, NO. 8, AUGUST 2015 11 cloud clustering-based instance segmentation for industrial bin- picking,

    Y. Xu, S. Arai, D. Liu, F. Lin, and K. Kosuge, “Fpcc: Fast point JOURNAL OF LATEX CLASS FILES, VOL. 14, NO. 8, AUGUST 2015 11 cloud clustering-based instance segmentation for industrial bin- picking,” Neurocomputing, vol. 494, pp. 255–268, 2022. 1

    2015

  15. [23]

    An integrated lidar-slam system for complex environment with noisy point clouds,

    K. Liu, “An integrated lidar-slam system for complex environment with noisy point clouds,”arXiv preprint arXiv:2212.05705, 2022. 1

    2022 arXiv

  16. [24]

    Deep learningfor3dpointclouds:Asurvey,

    Y.Guo,H.Wang,Q.Hu,H.Liu,L.Liu,andM.Bennamoun,“Deep learningfor3dpointclouds:Asurvey,” IEEEtransactionsonpattern analysis and machine intelligence, vol. 43, no. 12, pp. 4338–4364, 2020. 2

    2020

  17. [25]

    Density- aware chamfer distance as a comprehensive metric for point cloud completion,

    T. Wu, L. Pan, J. Zhang, T. Wang, Z. Liu, and D. Lin, “Density- aware chamfer distance as a comprehensive metric for point cloud completion,” inAdvances in Neural Information Processing Systems, vol. 34, 2021, pp. 29088–29100. 2, 3

    2021

  18. [26]

    Rethinking the compositionality of point clouds through regularization in the hyperbolic space,

    A. Montanaro, D. Valsesia, and E. Magli, “Rethinking the compositionality of point clouds through regularization in the hyperbolic space,” in Advances in Neural Information Processing Systems, A. H. Oh, A. Agarwal, D. Belgrave, and K. Cho, Eds., 2022. [Online]. Available: http...

    2022

  19. [27]

    Voxnet: A 3d convolutional neural network for real-time object recognition,

    D. Maturana and S. Scherer, “Voxnet: A 3d convolutional neural network for real-time object recognition,” in2015 IEEE/RSJ inter- national conference on intelligent robots and systems (IROS). IEEE, 2015, pp. 922–928. 2

    2015

  20. [28]

    Shape completion using 3d-encoder-predictor cnns and shape synthesis,

    A. Dai, C. Ruizhongtai Qi, and M. Nießner, “Shape completion using 3d-encoder-predictor cnns and shape synthesis,” inProceed- ings of the IEEE conference on computer vision and pattern recognition, 2017, pp. 5868–5877. 2

    2017

  21. [29]

    Pointgrid: A deep network for 3d shape understanding,

    T. Le and Y. Duan, “Pointgrid: A deep network for 3d shape understanding,” in Proceedings of the IEEE conference on computer vision and pattern recognition, 2018, pp. 9204–9214. 2

    2018

  22. [30]

    Voxel-based network for shape completion by leveraging edge generation,

    X. Wang, M. H. Ang, and G. H. Lee, “Voxel-based network for shape completion by leveraging edge generation,” inProceedings of the IEEE/CVF international conference on computer vision, 2021, pp. 13189–13198. 2

    2021

  23. [31]

    Pointnet: Deep learning onpointsetsfor3dclassificationandsegmentation,

    C. R. Qi, H. Su, K. Mo, and L. J. Guibas, “Pointnet: Deep learning onpointsetsfor3dclassificationandsegmentation,”in Proceedings oftheIEEEconferenceoncomputervisionandpatternrecognition ,2017, pp. 652–660. 2

    2017

  24. [32]

    Foldingnet: Point cloud auto-encoderviadeepgriddeformation,

    Y. Yang, C. Feng, Y. Shen, and D. Tian, “Foldingnet: Point cloud auto-encoderviadeepgriddeformation,”in ProceedingsoftheIEEE conference on computer vision and pattern recognition, 2018, pp. 206–

    2018

  25. [33]

    Detailpreservedpointcloudcom- pletion via separated feature aggregation,

    W.Zhang,Q.Yan,andC.Xiao,“Detailpreservedpointcloudcom- pletion via separated feature aggregation,” inEuropean Conference on Computer Vision. Springer, 2020, pp. 512–528. 2, 6

    2020

  26. [34]

    Cascaded refinement net- work for point cloud completion,

    X. Wang, M. H. Ang Jr, and G. H. Lee, “Cascaded refinement net- work for point cloud completion,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2020, pp. 790–799. 2, 6

    2020

  27. [35]

    A conditional point diffusion-refinement paradigm for 3d point cloud completion,

    Z. Lyu, Z. Kong, X. Xu, L. Pan, and D. Lin, “A conditional point diffusion-refinement paradigm for 3d point cloud completion,” arXiv preprint arXiv:2112.03530, 2021. 2, 3

    2021 arXiv

  28. [36]

    Deep unsupervised learning using nonequilibrium thermody- namics,

    J. Sohl-Dickstein, E. Weiss, N. Maheswaranathan, and S. Ganguli, “Deep unsupervised learning using nonequilibrium thermody- namics,” inInternational Conference on Machine Learning. PMLR, 2015, pp. 2256–2265. 2

    2015

  29. [37]

    Denoising diffusion probabilistic models,

    J. Ho, A. Jain, and P. Abbeel, “Denoising diffusion probabilistic models,” Advances in Neural Information Processing Systems, vol. 33, pp. 6840–6851, 2020. 2

    2020

  30. [38]

    3d shape generation and completion through point-voxel diffusion,

    L. Zhou, Y. Du, and J. Wu, “3d shape generation and completion through point-voxel diffusion,” in Proceedings of the IEEE/CVF International Conference on Computer Vision, 2021, pp. 5826–5835. 2

    2021

  31. [39]

    Diffusion probabilistic models for 3d point cloud generation,

    S. Luo and W. Hu, “Diffusion probabilistic models for 3d point cloud generation,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2021, pp. 2837–2845. 2, 10

    2021

  32. [40]

    P. J. Bickel and K. A. Doksum,Mathematical statistics: basic ideas and selected topics, volumes I-II package. Chapman and Hall/CRC, 2015. 2

    2015

  33. [41]

    Attention is all you need,

    A. Vaswani, N. Shazeer, N. Parmar, J. Uszkoreit, L. Jones, A. N. Gomez, Ł. Kaiser, and I. Polosukhin, “Attention is all you need,” Advances in neural information processing systems, vol. 30, 2017. 3

    2017

  34. [42]

    Point cloud completion by skip-attention network with hierarchical folding,

    X. Wen, T. Li, Z. Han, and Y.-S. Liu, “Point cloud completion by skip-attention network with hierarchical folding,” inCVPR, 2020. 3

    2020

  35. [43]

    Pu-gan: a point cloud upsampling adversarial network,

    R. Li, X. Li, C.-W. Fu, D. Cohen-Or, and P.-A. Heng, “Pu-gan: a point cloud upsampling adversarial network,” inProceedings of the IEEE/CVFinternationalconferenceoncomputervision ,2019,pp.7203–

    2019

  36. [44]

    Pu-gcn: Point cloud upsampling using graph convolutional networks,

    G.Qian,A.Abualshour,G.Li,A.Thabet,andB.Ghanem,“Pu-gcn: Point cloud upsampling using graph convolutional networks,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2021, pp. 11683–11692. 3, 7, 9

    2021

  37. [45]

    A point set generation network for 3d object reconstruction from a single image,

    H. Fan, H. Su, and L. J. Guibas, “A point set generation network for 3d object reconstruction from a single image,” inProceedings of the IEEE conference on computer vision and pattern recognition, 2017, pp. 605–613. 3, 10

    2017

  38. [46]

    A papier-mâché approach to learning 3d surface generation,

    T. Groueix, M. Fisher, V. G. Kim, B. C. Russell, and M. Aubry, “A papier-mâché approach to learning 3d surface generation,” in Proceedings of the IEEE conference on computer vision and pattern recognition, 2018, pp. 216–224. 3, 6, 10

    2018

  39. [47]

    3d shape reconstruction from 2d images with disentangled attribute flow,

    X. Wen, J. Zhou, Y.-S. Liu, H. Su, Z. Dong, and Z. Han, “3d shape reconstruction from 2d images with disentangled attribute flow,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2022, pp. 3803–3813. 3, 10

    2022

  40. [48]

    3d local features for direct pairwise registration,

    H. Deng, T. Birdal, and S. Ilic, “3d local features for direct pairwise registration,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2019, pp. 3244–3253. 3

    2019

  41. [49]

    Attention-based transfor- mation from latent features to point clouds,

    K. Zhang, X. Yang, Y. Wu, and C. Jin, “Attention-based transfor- mation from latent features to point clouds,” inProceedings of the AAAI Conference on Artificial Intelligence, vol. 36, no. 3, 2022, pp. 3291–3299. 3

    2022

  42. [50]

    Lake-net: topology- aware point cloud completion by localizing aligned keypoints,

    J. Tang, Z. Gong, R. Yi, Y. Xie, and L. Ma, “Lake-net: topology- aware point cloud completion by localizing aligned keypoints,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2022, pp. 1726–1735. 3

    2022

  43. [51]

    Morphing and sampling network for dense point cloud completion,

    M. Liu, L. Sheng, S. Yang, J. Shao, and S.-M. Hu, “Morphing and sampling network for dense point cloud completion,” in Proceedings of the AAAI conference on artificial intelligence, vol. 34, no. 07, 2020, pp. 11596–11603. 3

    2020

  44. [52]

    Learning representations and generative models for 3d point clouds,

    P.Achlioptas, O. Diamanti,I. Mitliagkas,and L.Guibas, “Learning representations and generative models for 3d point clouds,” in International conference on machine learning. PMLR, 2018, pp. 40–

    2018

  45. [53]

    Hyperbolic neural networks,

    O. Ganea, G. Bécigneul, and T. Hofmann, “Hyperbolic neural networks,”Advances in neural information processing systems, vol. 31,

  46. [54]

    Hyperbolic graph neural net- works,

    Q. Liu, M. Nickel, and D. Kiela, “Hyperbolic graph neural net- works,” Advances in neural information processing systems, vol. 32,

  47. [55]

    Poincaré mapsforanalyzingcomplexhierarchiesinsingle-celldata,

    A. Klimovskaia, D. Lopez-Paz, L. Bottou, and M. Nickel, “Poincaré mapsforanalyzingcomplexhierarchiesinsingle-celldata,” Nature communications, vol. 11, no. 1, p. 2966, 2020. 3

    2020

  48. [56]

    Hyperbolic deep neural networks: A survey,

    W. Peng, T. Varanka, A. Mostafa, H. Shi, and G. Zhao, “Hyperbolic deep neural networks: A survey,”arXiv preprint arXiv:2101.04562,

    arXiv

  49. [57]

    Geometric deep learning: going beyond euclidean data,

    M. M. Bronstein, J. Bruna, Y. LeCun, A. Szlam, and P. Van- dergheynst, “Geometric deep learning: going beyond euclidean data,” IEEE Signal Processing Magazine, vol. 34, no. 4, pp. 18–42,

  50. [58]

    Hyperbolic neural net- works++,

    R. Shimizu, Y. Mukuta, and T. Harada, “Hyperbolic neural net- works++,”arXiv preprint arXiv:2006.08210, 2020. 3

    2006 arXiv

  51. [59]

    Poincaré embeddings for learning hier- archical representations,

    M. Nickel and D. Kiela, “Poincaré embeddings for learning hier- archical representations,”Advances in neural information processing systems, vol. 30, 2017. 3

    2017

  52. [60]

    Learning continuous hierarchies in the lorentz model of hy- perbolic geometry,

    ——, “Learning continuous hierarchies in the lorentz model of hy- perbolic geometry,” inInternational Conference on Machine Learning. PMLR, 2018, pp. 3779–3788. 3

    2018

  53. [61]

    Unsupervised discovery of the long-tail in instance segmentation using hierar- chical self-supervision,

    Z. Weng, M. G. Ogut, S. Limonchik, and S. Yeung, “Unsupervised discovery of the long-tail in instance segmentation using hierar- chical self-supervision,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2021, pp. 2603–2612. 3

    2021

  54. [62]

    Hyperbolic image segmentation,

    M. G. Atigh, J. Schoep, E. Acar, N. van Noord, and P. Mettes, “Hyperbolic image segmentation,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2022, pp. 4453–4462. 3

    2022

  55. [63]

    Hyperbolic image embeddings,

    V. Khrulkov, L. Mirvakhabova, E. Ustinova, I. Oseledets, and V. Lempitsky, “Hyperbolic image embeddings,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2020, pp. 6418–6428. 3

    2020

  56. [64]

    Hy- perbolic visual embedding learning for zero-shot recognition,

    S.Liu,J.Chen,L.Pan,C.-W.Ngo,T.-S.Chua,andY.-G.Jiang,“Hy- perbolic visual embedding learning for zero-shot recognition,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2020, pp. 9273–9281. 3 JOURNAL OF LATEX CLASS FILES, VOL. 14, NO. 8, AU...

    2020

  57. [65]

    Hyperbolic vision transformers: Combining improve- mentsinmetriclearning,

    A. Ermolov, L. Mirvakhabova, V. Khrulkov, N. Sebe, and I. Os- eledets, “Hyperbolic vision transformers: Combining improve- mentsinmetriclearning,”in ProceedingsoftheIEEE/CVFConference on Computer Vision and Pattern Recognition, 2022, pp. 7409–7419. 3

    2022

  58. [66]

    Hyperbolic deep neural networks: A survey,

    W. Peng, T. Varanka, A. Mostafa, H. Shi, and G. Zhao, “Hyperbolic deep neural networks: A survey,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 44, no. 12, pp. 10023–10044,

  59. [67]

    Why gradient clip- ping accelerates training: A theoretical justification for adaptivity,

    J. Zhang, T. He, S. Sra, and A. Jadbabaie, “Why gradient clip- ping accelerates training: A theoretical justification for adaptivity,” arXiv preprint arXiv:1905.11881, 2019. 4

    1905 arXiv

  60. [68]

    Hyperbolic trigonometry and its application in the poincaré ball model of hyperbolic geometry,

    A. A. Ungar, “Hyperbolic trigonometry and its application in the poincaré ball model of hyperbolic geometry,” Computers & Mathematics with Applications, vol. 41, no. 1-2, pp. 135–147, 2001. 4

    2001

  61. [69]

    World Scientific, 2008

    ——, Analytic hyperbolic geometry and Albert Einstein’s special theory of relativity. World Scientific, 2008. 4

    2008

  62. [70]

    A gyrovector space approach to hyperbolic geometry,

    ——, “A gyrovector space approach to hyperbolic geometry,” Synthesis Lectures on Mathematics and Statistics, vol. 1, no. 1, pp. 1–194, 2008. 4

    2008

  63. [71]

    Hyperbolic con- trastive learning,

    Y. Yue, F. Lin, K. D. Yamada, and Z. Zhang, “Hyperbolic con- trastive learning,”arXiv preprint arXiv:2302.01409, 2023. 4

    2023 arXiv

  64. [72]

    Geoopt: Riemannian optimization in pytorch,

    M. Kochurov, R. Karimov, and S. Kozlukov, “Geoopt: Riemannian optimization in pytorch,” 2020. 4

    2020

  65. [73]

    Topnet: Structural point cloud decoder,

    L. P. Tchapmi, V. Kosaraju, H. Rezatofighi, I. Reid, and S. Savarese, “Topnet: Structural point cloud decoder,” inCVPR, 2019. 6, 8

    2019

  66. [74]

    Grnet: Gridding residual network for dense point cloud completion,

    H. Xie, H. Yao, S. Zhou, J. Mao, S. Zhang, and W. Sun, “Grnet: Gridding residual network for dense point cloud completion,” in ECCV, 2020. 6, 8

    2020

  67. [75]

    Fbnet: Feedback network for point cloud completion,

    X. Yan, H. Yan, J. Wang, H. Du, Z. Wu, D. Xie, S. Pu, and L. Lu, “Fbnet: Feedback network for point cloud completion,” in European Conference on Computer Vision. Springer, 2022, pp. 676–

    2022

  68. [76]

    Pmp-net: Point cloud completion by learning multi-step point moving paths,

    X. Wen, P. Xiang, Z. Han, Y.-P. Cao, P. Wan, W. Zheng, and Y.- S. Liu, “Pmp-net: Point cloud completion by learning multi-step point moving paths,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2021, pp. 7443–7452. 6

    2021

  69. [77]

    Pf-net: Point fractal network for 3d point cloud completion,

    Z. Huang, Y. Yu, J. Xu, F. Ni, and X. Le, “Pf-net: Point fractal network for 3d point cloud completion,” inCVPR, 2020. 6, 8

    2020

  70. [78]

    A scalable active framework for region annotation in 3d shape collections,

    L. Yi, V. G. Kim, D. Ceylan, I.-C. Shen, M. Yan, H. Su, C. Lu, Q. Huang, A. Sheffer, and L. Guibas, “A scalable active framework for region annotation in 3d shape collections,”ACM Transactions on Graphics (ToG), vol. 35, no. 6, pp. 1–12, 2016. 5

    2016

  71. [79]

    Shapenet: An information-rich 3d model repository,

    A. X. Chang, T. Funkhouser, L. Guibas, P. Hanrahan, Q. Huang, Z. Li, S. Savarese, M. Savva, S. Song, H. Su et al., “Shapenet: An information-rich 3d model repository,” arXiv preprint arXiv:1512.03012, 2015. 5, 6, 7

    2015 arXiv

  72. [80]

    Vision meets robotics: The kitti dataset,

    A. Geiger, P. Lenz, C. Stiller, and R. Urtasun, “Vision meets robotics: The kitti dataset,” The International Journal of Robotics Research, vol. 32, no. 11, pp. 1231–1237, 2013. 6

    2013

  73. [81]

    Snowflake point deconvolution for point cloud comple- tion and generation with skip-transformer,

    P. Xiang, X. Wen, Y.-S. Liu, Y.-P. Cao, P. Wan, W. Zheng, and Z. Han, “Snowflake point deconvolution for point cloud comple- tion and generation with skip-transformer,”IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 45, no. 5, pp. 6320– 6338, 2022. 6

    2022

  74. [82]

    Cosmospropa- gation network: Deep learning model for point cloud completion,

    F.Lin,Y.Xu,Z.Zhang,C.Gao,andK.D.Yamada,“Cosmospropa- gation network: Deep learning model for point cloud completion,” Neurocomputing, vol. 507, pp. 221–234, 2022. 7, 9

    2022

  75. [83]

    Pytorch: An im- perative style, high-performance deep learning library,

    A. Paszke, S. Gross, F. Massa, A. Lerer, J. Bradbury, G. Chanan, T. Killeen, Z. Lin, N. Gimelshein, L. Antigaet al., “Pytorch: An im- perative style, high-performance deep learning library,”Advances in neural information processing systems, vol. 32, 2019. 7

    2019

  76. [84]

    Patch-based progressive 3d point set upsampling,

    W.Yifan,S.Wu,H.Huang,D.Cohen-Or,andO.Sorkine-Hornung, “Patch-based progressive 3d point set upsampling,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recogni- tion, 2019, pp. 5958–5967. 7, 9

    2019

  77. [85]

    Pointcloudupsampling via cascaded refinement network,

    H.Du,X.Yan,J.Wang,D.Xie,andS.Pu,“Pointcloudupsampling via cascaded refinement network,” in Proceedings of the Asian Conference on Computer Vision, 2022, pp. 586–601. 7, 9

    2022

  78. [86]

    What do single-view 3d reconstruction networks learn?

    M. Tatarchenko, S. R. Richter, R. Ranftl, Z. Li, V. Koltun, and T. Brox, “What do single-view 3d reconstruction networks learn?” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2019, pp. 3405–3414. 8

    2019

  79. [87]

    3d-r2n2: A unified approach for single and multi-view 3d object reconstruc- tion,

    C. B. Choy, D. Xu, J. Gwak, K. Chen, and S. Savarese, “3d-r2n2: A unified approach for single and multi-view 3d object reconstruc- tion,” in Computer Vision–ECCV 2016: 14th European Conference, Amsterdam, The Netherlands, October 11-14, 2016, Proceedings, Part VIII 14. Springe...

    2016

  80. [88]

    Pixel2mesh:Generating3dmeshmodelsfromsinglergbimages,

    N. Wang, Y. Zhang, Z. Li, Y. Fu, W. Liu, and Y.-G. Jiang, “Pixel2mesh:Generating3dmeshmodelsfromsinglergbimages,” in Proceedings of the European conference on computer vision (ECCV), 2018, pp. 52–67. 10

    2018

  81. [89]

    Occupancy networks: Learning 3d reconstruction in function space,

    L. Mescheder, M. Oechsle, M. Niemeyer, S. Nowozin, and A. Geiger, “Occupancy networks: Learning 3d reconstruction in function space,” inProceedings of the IEEE/CVF conference on com- puter vision and pattern recognition, 2019, pp. 4460–4470. 10

    2019

  82. [90]

    Pix2vox: Context- aware 3d reconstruction from single and multi-view images,

    H. Xie, H. Yao, X. Sun, S. Zhou, and S. Zhang, “Pix2vox: Context- aware 3d reconstruction from single and multi-view images,” in Proceedings of the IEEE/CVF international conference on computer vision, 2019, pp. 2690–2698. 10

    2019

  83. [91]

    Pix2vox++: Multi-scalecontext-aware3dobjectreconstructionfromsingleand multiple images,

    H. Xie, H. Yao, S. Zhang, S. Zhou, and W. Sun, “Pix2vox++: Multi-scalecontext-aware3dobjectreconstructionfromsingleand multiple images,”International Journal of Computer Vision, vol. 128, no. 12, pp. 2919–2935, 2020. 10 Fangzhou Lin received his M.S. and Ph.D. in in- formation...

    2020

Pith tools

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