Pith. sign in

REVIEW 3 major objections 5 minor 1 cited by

PointCFormer: a Relation-based Progressive Feature Extraction Network for Point Cloud Completion

T0 review · 3 major / 5 minor · reviewed 2026-08-11 · deepseek-v4-flash

Pith's one-line read PointCFormer improves point cloud completion by weighting each neighbor's contribution with spatial and feature relation metrics.

desk verdict Competent incremental point-cloud-completion paper with a plausible synthetic-benchmark story, undermined by an implausible KITTI fidelity number and two unreported hyperparameters. read the letter →

arxiv 2412.08421 v2 pith:QUMTYVER submitted 2024-12-11 cs.CV

classification cs.CV
keywords pointcloudcompletiontransformerlocalgeometricrelationshipperceptionk-nearestneighborweightingprogressivefeatureextractionproxycorrectionChamferdistance3Dshapereconstruction
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

PointCFormer is a transformer-based network for reconstructing a full 3D shape from an incomplete point cloud. The paper's central claim is that feature extraction improves when each neighbor's contribution to a target point is explicitly weighted by how related the two points are, measured both by spatial position and by learned feature vectors. To avoid losing global shape information during down-sampling, the network alternates global self-attention queries with a scale-tailored local feature extractor, gradually moving from dense coarse features to sparse refined ones. A correction module then re-injects dense input information into the latent point proxies. On PCN, ShapeNet-55, ShapeNet-34/Unseen-21, and KITTI, the reported Chamfer distances and F-Score are the best among the compared methods, with the largest ablation gain coming from the relation-based local perception.

What carries the argument

The central mechanism is the relation-based local geometric relationship perception module, which turns a standard k-NN neighborhood aggregation into a weighted aggregation. Its two metrics are $R_1(Q,V_i) = (|x_Q-x_{V_i}|, |y_Q-y_{V_i}|, |z_Q-z_{V_i}|)$, the component-wise Manhattan displacement in 3D, and $R_2(Q,V_i) = \left|\frac{1}{M}\sum_{j=1}^M \mathrm{vec}(Q,V_j) - \mathrm{vec}(Q,V_i)\right|$, the deviation of a neighbor's directed feature edge from the average directed edge of a small subset of neighbors. After concatenation, an MLP maps the pair to a per-neighbor contribution weight used to modulate the EdgeConv features before max-pooling. This weighted local perception is embedded in a progressive feature extractor that alternates global self-attention with scale-tailored local feature extraction, and in the point proxy correction module that pulls the latent proxies back toward the dense input distribution.

What would settle it

Run the released code on PCN and ShapeNet-55 while sweeping $M$ over a range such as 1, 4, 8, and 16 and $\lambda$ over values including 0, 0.1, 1, and 10. If the reported CD-$\ell_1$ of 6.41 on PCN and the margin over AdaPoinTr appear only at a single untested-in-public configuration, or if setting $\lambda = 0$ or changing $M$ alters the ranking, the claim that the relation-weighted architecture drives the gain is not supported. A second check is to replace the average-edge term in $R_2$ with a fixed geometric baseline such as the local centroid and measure whether the advantage disappears.

Watch

Extended reading notes

Core claim

The discovery the paper argues for is that k-nearest-neighbor pooling is a bottleneck in point cloud completion and can be replaced by a learned weighting scheme. For each target point, the method computes a spatial relation metric $R_1$ (Manhattan distances to each neighbor) and a feature-space relation metric $R_2$ (the absolute difference between a neighbor's directed edge and the average directed edge of a small subset of neighbors), concatenates them, and uses an MLP to produce a contribution weight per neighbor before max-pooling. This relation-based local geometric relationship perception is inserted into a progressive feature extractor that alternates self-attention and down-sampling, and into a correction module that refines the latent point proxies. The paper reports that these components together bring average Chamfer distance on PCN to 6.41 (CD-$\ell_1$, multiplied by $10^3$) with F-Score@1\% of 0.855, and ShapeNet-55 CD-$\ell_2$ to 0.73, both better than the compared methods. The ablation attributes the largest part of the gain to the relation-based local perception module rather than to the global attention or correction modules.

Load-bearing premise

The load-bearing premise is that the reported gains come from the architecture itself rather than from the choice of two unreported hyperparameters: the subset size $M$ used in the feature-relation metric $R_2$ and the weight $\lambda$ on the denoising auxiliary loss. If those values were selected after seeing test results, the benchmark improvements could be artifacts of model selection rather than of the proposed modules.

Editorial extensions

If this is right

  • If the central claim holds, k-NN-based feature extractors in other point cloud tasks can be improved by replacing unweighted pooling with relation-weighted aggregation, since the learned weighting is not specific to completion.
  • The reported benchmark results imply that one relation-aware extractor can simultaneously improve global shape fidelity, measured by Chamfer distance, and local detail, measured by F-Score, against AdaPoinTr, 3DMambaComplete, and PointAttN on PCN, ShapeNet-55, and ShapeNet-34.
  • The progressive down-sampling schedule lets dense input information be retained early and refined gradually, which the paper claims avoids both the global-structure loss of aggressive down-sampling and the high cost of a purely attention-based extractor.
  • The correction module implies that latent point proxies are more accurate when dense input points are explicitly re-introduced in a low-dimensional space rather than relying only on cross-attention with the encoder output.

Reading between the lines

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

  • I would read the ablation as evidence that the relation weighting is the most transferable piece: the same module produces gains in both the extractor and the correction module, so it could be plugged into other completion backbones without redesigning the whole extractor.
  • Because the paper does not report the value of the subset size $M$ in $R_2$ or the weight $\lambda$ on the denoising loss, a reader should not assume the reported gains are invariant to those settings until a sensitivity sweep is published.
  • Since $R_2$ subtracts an average edge direction, it behaves like a local residual estimate; a natural testable variant would replace the average with a local PCA tangent direction and compare completion quality.
  • The two-phase progressive extractor suggests a general encoder design for downstream tasks such as segmentation or detection, where the same tension between global context and local detail appears.
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

3 major / 5 minor

Summary. The paper presents PointCFormer, a transformer-based network for point cloud completion. It introduces three modules: a local geometric relationship perception (LGRP) module that weights kNN neighbors via R1 (spatial Manhattan distance) and R2 (relative directed-edge deviation) metrics; a progressive feature extractor (PFE) that alternates self-attention and scale-tailored local feature extraction (ST-LFE) while downsampling; and a point proxy correction module (CM) that re-injects dense input information into latent point proxies. The method is trained on PCN, ShapeNet-55, ShapeNet-34/21, and KITTI, and the main tables report state-of-the-art Chamfer distances (CD-l1 6.41 on PCN, CD-l2 0.73 on ShapeNet-55) and favorable F-scores compared with AdaPoinTr, 3DMambaComplete, and others. An ablation on PCN shows that each module contributes to the final score.

Significance. If the reported results are accurate, the paper offers a practically useful architecture that improves completion quality across multiple benchmarks by a small but consistent margin (e.g., CD-l1 6.41 vs 6.53 for AdaPoinTr on PCN), and it provides a thorough module analysis and released code. The relation-based weighting idea is well motivated, and the progressive extractor's combination of global and local queries is a reasonable design. However, the empirical claim currently rests on at least one implausible result — the KITTI Fidelity of 0.001 — and on two unreported hyperparameters (M and lambda), which makes the significance conditional on those points being resolved.

major comments (3)
  1. [Table 4 / Experiments] Table 4 reports a KITTI Fidelity of 0.001 for PointCFormer versus 0.237 for AdaPoinTr, under a header stating values are multiplied by 1000. This implies an actual fidelity around 1e-6, i.e., nearly every completed point coincides with a ground-truth LiDAR point. Given the KITTI ground truth is itself incomplete and noisy, and the MMD improvement is only from 0.392 to 0.353, this value is not plausible under the stated protocol. The paper provides no explanation of any altered evaluation protocol. If this number is a typo or a unit error, the KITTI state-of-the-art claim is unsupported and the reliability of the other benchmark tables is cast in doubt because they are produced by the same pipeline. Please re-check the KITTI evaluation and report the corrected value or the exact protocol.
  2. [Eq. (2) and Network Optimization] The definition of R2 in Eq. (2) depends on a subset size M ('we select a small subset (M points) of the nearest neighbors'), and the final objective J_PC = J0 + J1 + λ Jdenoise depends on the weight λ. Neither M nor λ is reported in the main text or in the supplementary 'Hyper-parameter Recommendation' section, which states k=16 but is silent on M and λ. Without these values, the method cannot be reproduced, and the reported improvements could have resulted from fine-tuning M and λ on the test benchmarks. Please report the chosen values and provide a sensitivity analysis over reasonable ranges.
  3. [Table 1] The PointAttN(aaai2024) row in Table 1 is incomplete: values for Sofa, Table, Boat, and F-Score@1% are missing, and the Cabinet value 17.923 is far outside the range of all other methods (the other methods range from 8.34 to 10.47 on that category, except for PCN at 22.70). This incomplete row makes the 'leads in almost all metrics' claim hard to verify. Please provide the full PointAttN results as published in its original paper or clearly state why some entries are unavailable.
minor comments (5)
  1. [Supplementary Material] The supplementary material begins with 'Anonymous submission' and contains placeholder citations such as '(?)' in 'Similar to the AdapoinTr (?) protocol' and 'Projected-ShapeNet-55 (?)'. These need to be cleaned up before publication.
  2. [References / Supplementary] The main text cites 'Zhong et al. 2024' (the same arXiv preprint) for a more detailed analysis of R1 and R2, but the supplementary material appears to be part of this submission, not an external reference. This self-referential citation is confusing and should be replaced with a cross-reference to the supplement within this paper.
  3. [Supplementary Tables 4-5] The supplementary tables report 'PointCFormer-Plus' results without any definition of this variant. If Plus is a larger model or an ensemble, define it in the main paper or remove the entries.
  4. [Experiments] The paper reports a single run for each method and no error bars or multiple seeds. Given that the reported margins over the closest competitor are sometimes small (e.g., 0.73 vs 0.77 CD-l2 on ShapeNet-55), a statement on variance or a multi-seed evaluation would strengthen the SOTA claim.
  5. [Table 2] In Table 2, the F-Score@1% of PointCFormer (0.499) is slightly lower than AdaPoinTr (0.503); the text acknowledges this but could state it more explicitly to avoid overclaiming.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: SOTA claims rest on external benchmarks; the only self-citation is a non-load-bearing pointer to the paper's own supplementary.

full rationale

The paper's central claim is benchmark performance against pre-existing external datasets (PCN, ShapeNet-55, ShapeNet-34/Unseen-21, KITTI) and previously published methods. The training objectives in Eqs. (3)-(5) are standard Chamfer-distance and denoising losses; none of these equations encode the reported rankings or F-Score values. The proposed modules (LGRP, progressive feature extractor, correction module) are architectural components whose contributions are assessed through ablations on the same external metrics, not through quantities defined in terms of those metrics. The relation metrics R1 and R2 are learned feature-weighting mechanisms, not definitions of the evaluation criteria. The unreported subset size M in Eq. (2) and the unreported weight lambda in the final objective J_PC are reproducibility/model-selection concerns, but there is no evidence in the paper that these were fitted to the test rankings, so they do not constitute a circular reduction. The KITTI Fidelity value of 0.001 in Table 4 is a plausible metric bug or typo, but that is an empirical correctness risk, not a form of circular reasoning. The only self-referential element is the citation 'Supplementary Material(Zhong et al. 2024)' and the corresponding reference-list entry, which points to the paper's own supplementary material; this is not load-bearing for any central claim and does not force any result. Overall, the derivation chain is self-contained against external benchmarks, and no prediction reduces by construction to its inputs.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The central claim rests on standard deep learning practices: benchmark datasets, metric choices, and architecture hyperparameters. No new physical or mathematical entities are introduced. The most fragile inputs are the unreported M and lambda, which directly enter two formulas, and the design assumption that kNN in Euclidean space captures local relevance.

free parameters (3)
  • k, kNN neighborhood size = 16
    Used in LGRP and ST-LFE; selected by a validation sweep in Supplemental Table 1, where k=8/16/24 are compared.
  • M, subset size for R2 average
    Appears in Eq. (2) as the number of nearest neighbors averaged to form the primary change trend; never specified in main text or supplement.
  • lambda, denoising loss weight
    Appears in J_PC = J0 + J1 + lambda * Jdenoise; no value reported.
assumptions (4)
  • domain assumption Euclidean kNN identifies the locally relevant neighborhood for completion, and learned R1/R2 weights can filter out irrelevant neighbors.
    The LGRP module assumes spatial proximity in input coordinates is a useful proxy for geometric relevance; introduced in 'Local Geometric Relationship Perception'.
  • domain assumption Chamfer distance and F-Score@1% are sufficient and reliable metrics for completion quality.
    All benchmark comparisons use these metrics from prior work; the paper does not analyze failure modes or metric sensitivity.
  • domain assumption Progressive FPS downsampling combined with alternating global and local queries preserves global structure without substantial information loss.
    This is the core design premise of the progressive feature extractor; no information-theoretic or empirical loss analysis is provided.
  • domain assumption Reported benchmark protocols for PCN, ShapeNet-55/34, and KITTI are followed as in prior work.
    The paper states 'we adhere to the standard protocols'; it does not include dataset splits or preprocessing code.

how reviews work

0 comments
Cite this review

Pith. "Pith review of PointCFormer: a Relation-based Progressive Feature Extraction Network for Point Cloud Completion." pith.science (2026). https://pith.science/paper/QUMTYVER

@misc{pith2026241208421,
  author       = {Pith},
  title        = {Pith review of: PointCFormer: a Relation-based Progressive Feature Extraction Network for Point Cloud Completion},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/QUMTYVER}},
  note         = {Machine review of arXiv:2412.08421}
}
read the original abstract

Point cloud completion aims to reconstruct the complete 3D shape from incomplete point clouds, and it is crucial for tasks such as 3D object detection and segmentation. Despite the continuous advances in point cloud analysis techniques, feature extraction methods are still confronted with apparent limitations. The sparse sampling of point clouds, used as inputs in most methods, often results in a certain loss of global structure information. Meanwhile, traditional local feature extraction methods usually struggle to capture the intricate geometric details. To overcome these drawbacks, we introduce PointCFormer, a transformer framework optimized for robust global retention and precise local detail capture in point cloud completion. This framework embraces several key advantages. First, we propose a relation-based local feature extraction method to perceive local delicate geometry characteristics. This approach establishes a fine-grained relationship metric between the target point and its k-nearest neighbors, quantifying each neighboring point's contribution to the target point's local features. Secondly, we introduce a progressive feature extractor that integrates our local feature perception method with self-attention. Starting with a denser sampling of points as input, it iteratively queries long-distance global dependencies and local neighborhood relationships. This extractor maintains enhanced global structure and refined local details, without generating substantial computational overhead. Additionally, we develop a correction module after generating point proxies in the latent space to reintroduce denser information from the input points, enhancing the representation capability of the point proxies. PointCFormer demonstrates state-of-the-art performance on several widely used benchmarks. Our code is available at https://github.com/Zyyyyy0926/PointCFormer_Plus_Pytorch.

Figures

Figures reproduced from arXiv: 2412.08421 by the authors.

Figure 1
Figure 1. Improved global shape and enriched local detail in [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Overview of PointCFormer framework: Initially, we extract representative sampling points and their local features [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Comparison of local feature extraction: traditional [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗
Figures from the paper (7 more)
Figure 6
Figure 6. Figure 6: Scale-tailored local feature extractor. where vec represents the vector of differences (directed edges) between two points [PITH_FULL_IMAGE:figures/full_fig_p004_6.png]
Figure 7
Figure 7. Figure 7: Visual examples of point cloud completion results on the PCN dataset using different methods. [PITH_FULL_IMAGE:figures/full_fig_p006_7.png]
Figure 8
Figure 8. Figure 8: Qualitative results on KITTI: dual-view represen [PITH_FULL_IMAGE:figures/full_fig_p006_8.png]
Figure 1
Figure 1. Figure 1: The first row of the figure presents the training curves [PITH_FULL_IMAGE:figures/full_fig_p011_1.png]
Figure 2
Figure 2. Figure 2: More qualitative results on PCN [PITH_FULL_IMAGE:figures/full_fig_p014_2.png]
Figure 3
Figure 3. Figure 3: More qualitative results on ShapeNet-55. [PITH_FULL_IMAGE:figures/full_fig_p015_3.png]
Figure 4
Figure 4. Figure 4: More qualitative results on KITTI [PITH_FULL_IMAGE:figures/full_fig_p016_4.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

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

  1. WLTCL: Wide Field-of-View 3-D LiDAR Truck Compartment Automatic Localization System

    cs.CV 2025-04 conditional novelty 5.0 of 10

    A rotating LiDAR system with line fusion and contour completion localizes fence truck compartment corners with mean relative errors of 1.9% to 6.2% depending on truck size.

Reference graph

Works this paper leans on

40 extracted references · 25 canonical work pages · cited by 1 Pith paper

  1. [1]

    , " * write output.state after.block = add.period write newline

    ENTRY address archivePrefix author booktitle chapter edition editor eid eprint howpublished institution isbn journal key month note number organization pages publisher school series title type volume year label extra.label sort.label short.list INTEGERS output.state before.all mid.sentence after.sentence after.block FUNCTION init.state.consts #0 'before.a...

  2. [2]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION word.in bbl.in capitalize " " * FUNCT...

  3. [3]

    Bi, Z.; and Wang, L. 2010. Advances in 3D data acquisition and processing for industrial applications. Robotics and Computer-Integrated Manufacturing, 26(5): 403--413

  4. [4]

    Chen, C.; and Yang, B. 2016. Dynamic occlusion detection and inpainting of in situ captured terrestrial laser scanning point clouds sequence. ISPRS Journal of Photogrammetry and Remote Sensing, 119: 90--107

  5. [5]

    Chen, Z.; Long, F.; Qiu, Z.; Yao, T.; Zhou, W.; Luo, J.; and Mei, T. 2023. Anchorformer: Point cloud completion from discriminative nodes. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 13581--13590

  6. [6]

    Dai, A.; Ruizhongtai Qi, C.; and Nie ner, M. 2017. Shape completion using 3d-encoder-predictor cnns and shape synthesis. In Proceedings of the IEEE conference on computer vision and pattern recognition, 5868--5877

  7. [7]

    Dosovitskiy, A.; Beyer, L.; Kolesnikov, A.; Weissenborn, D.; Zhai, X.; Unterthiner, T.; Dehghani, M.; Minderer, M.; Heigold, G.; Gelly, S.; et al. 2020. An image is worth 16x16 words: Transformers for image recognition at scale. arXiv preprint arXiv:2010.11929

  8. [8]

    Fei, B.; Yang, W.; Ma, L.; and Chen, W.-M. 2023. DcTr: Noise-robust point cloud completion by dual-channel transformer with cross-attention. Pattern Recognition, 133: 109051

Show all 40 references
  1. [9]

    Gu, A.; and Dao, T. 2023. Mamba: Linear-Time Sequence Modeling with Selective State Spaces. arXiv preprint arXiv:2312.00752

  2. [10]

    R.; and Hu, S.-M

    Guo, M.-H.; Cai, J.-X.; Liu, Z.-N.; Mu, T.-J.; Martin, R. R.; and Hu, S.-M. 2021. Pct: Point cloud transformer. Computational Visual Media, 7: 187--199

  3. [11]

    Han, K.; Wang, Y.; Chen, H.; Chen, X.; Guo, J.; Liu, Z.; Tang, Y.; Xiao, A.; Xu, C.; Xu, Y.; et al. 2022. A survey on vision transformer. IEEE transactions on pattern analysis and machine intelligence, 45(1): 87--110

  4. [12]

    Han, X.; Li, Z.; Huang, H.; Kalogerakis, E.; and Yu, Y. 2017. High-resolution shape completion using deep neural networks for global structure and local geometry inference. In Proceedings of the IEEE international conference on computer vision, 85--93

  5. [13]

    LeCun, Y.; Bottou, L.; Bengio, Y.; and Haffner, P. 1998. Gradient-based learning applied to document recognition. Proceedings of the IEEE, 86(11): 2278--2324

  6. [14]

    Li, Y.; Yang, W.; and Fei, B. 2024. 3dmambacomplete: Exploring structured state space model for point cloud completion. arXiv preprint arXiv:2404.07106

  7. [15]

    Mao, A.; Yan, B.; Ma, Z.; and He, Y. 2024. Denoising Point Clouds in Latent Space via Graph Convolution and Invertible Neural Network. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 5768--5777

  8. [16]

    Mao, J.; Xue, Y.; Niu, M.; Bai, H.; Feng, J.; Liang, X.; Xu, H.; and Xu, C. 2021. Voxel transformer for 3d object detection. In Proceedings of the IEEE/CVF international conference on computer vision, 3164--3173

  9. [17]

    Misra, I.; Girdhar, R.; and Joulin, A. 2021. An end-to-end transformer model for 3d object detection. In Proceedings of the IEEE/CVF international conference on computer vision, 2906--2917

  10. [18]

    E.; and Huang, G

    Pan, X.; Xia, Z.; Song, S.; Li, L. E.; and Huang, G. 2021. 3d object detection with pointformer. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 7463--7472

  11. [19]

    R.; Su, H.; Mo, K.; and Guibas, L

    Qi, C. R.; Su, H.; Mo, K.; and Guibas, L. J. 2017 a . Pointnet: Deep learning on point sets for 3d classification and segmentation. In Proceedings of the IEEE conference on computer vision and pattern recognition, 652--660

  12. [20]

    R.; Yi, L.; Su, H.; and Guibas, L

    Qi, C. R.; Yi, L.; Su, H.; and Guibas, L. J. 2017 b . Pointnet++: Deep hierarchical feature learning on point sets in a metric space. Advances in neural information processing systems, 30

  13. [21]

    Stutz, D.; and Geiger, A. 2018. Learning 3d shape completion from laser scan data with weak supervision. In Proceedings of the IEEE conference on computer vision and pattern recognition, 1955--1964

  14. [22]

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

  15. [23]

    Touvron, H.; Cord, M.; Douze, M.; Massa, F.; Sablayrolles, A.; and J \'e gou, H. 2021. Training data-efficient image transformers & distillation through attention. In International conference on machine learning, 10347--10357. PMLR

  16. [24]

    N.; Kaiser, .; and Polosukhin, I

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

  17. [25]

    Wang, J.; Cui, Y.; Guo, D.; Li, J.; Liu, Q.; and Shen, C. 2024. PointAttN: You Only Need Attention for Point Cloud Completion. Proceedings of the AAAI Conference on Artificial Intelligence, 38(6): 5472--5480

  18. [26]

    Wang, P.-S. 2023. Octformer: Octree-based transformers for 3d point clouds. ACM Transactions on Graphics (TOG), 42(4): 1--11

  19. [27]

    E.; Bronstein, M

    Wang, Y.; Sun, Y.; Liu, Z.; Sarma, S. E.; Bronstein, M. M.; and Solomon, J. M. 2019. Dynamic graph cnn for learning on point clouds. ACM Transactions on Graphics (tog), 38(5): 1--12

  20. [28]

    J.; Navab, N.; and Tombari, F

    Wang, Y.; Tan, D. J.; Navab, N.; and Tombari, F. 2022. Learning local displacements for point cloud completion. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 1568--1577

  21. [29]

    Wen, X.; Xiang, P.; Han, Z.; Cao, Y.-P.; Wan, P.; Zheng, W.; and Liu, Y.-S. 2022. Pmp-net++: Point cloud completion by transformer-enhanced multi-step point moving paths. IEEE Transactions on Pattern Analysis and Machine Intelligence, 45(1): 852--867

  22. [30]

    Wu, Z.; Song, S.; Khosla, A.; Yu, F.; Zhang, L.; Tang, X.; and Xiao, J. 2015. 3d shapenets: A deep representation for volumetric shapes. In Proceedings of the IEEE conference on computer vision and pattern recognition, 1912--1920

  23. [31]

    Xiang, P.; Wen, X.; Liu, Y.-S.; Cao, Y.-P.; Wan, P.; Zheng, W.; and Han, Z. 2021. Snowflakenet: Point cloud completion by snowflake point deconvolution with skip-transformer. In Proceedings of the IEEE/CVF international conference on computer vision, 5499--5509

  24. [32]

    Xie, H.; Yao, H.; Zhou, S.; Mao, J.; Zhang, S.; and Sun, W. 2020. Grnet: Gridding residual network for dense point cloud completion. In European conference on computer vision, 365--381. Springer

  25. [33]

    Yan, X.; Yan, H.; Wang, J.; Du, H.; Wu, Z.; Xie, D.; Pu, S.; and Lu, L. 2022. Fbnet: Feedback network for point cloud completion. In European Conference on Computer Vision, 676--693. Springer

  26. [34]

    Yang, Y.; Feng, C.; Shen, Y.; and Tian, D. 2018. Foldingnet: Point cloud auto-encoder via deep grid deformation. In Proceedings of the IEEE conference on computer vision and pattern recognition, 206--215

  27. [35]

    Yu, X.; Rao, Y.; Wang, Z.; Liu, Z.; Lu, J.; and Zhou, J. 2021. Pointr: Diverse point cloud completion with geometry-aware transformers. In Proceedings of the IEEE/CVF international conference on computer vision, 12498--12507

  28. [36]

    Yu, X.; Rao, Y.; Wang, Z.; Lu, J.; and Zhou, J. 2023. AdaPoinTr: Diverse Point Cloud Completion With Adaptive Geometry-Aware Transformers. IEEE Transactions on Pattern Analysis and Machine Intelligence, 45(12): 14114--14130

  29. [37]

    Yuan, W.; Khot, T.; Held, D.; Mertz, C.; and Hebert, M. 2018. Pcn: Point completion network. In 2018 international conference on 3D vision (3DV), 728--737. IEEE

  30. [38]

    Zhang, W.; Zhou, H.; Dong, Z.; Liu, J.; Yan, Q.; and Xiao, C. 2022. Point cloud completion via skeleton-detail transformer. IEEE Transactions on Visualization and Computer Graphics, 29(10): 4229--4242

  31. [39]

    Zhong, Y.; Quan, W.; Yan, D.-m.; Jiang, J.; and Wei, Y. 2024. PointCFormer: a Relation-based Progressive Feature Extraction Network for Point Cloud Completion. arXiv preprint arXiv:2412.08421

  32. [40]

    Zhou, H.; Cao, Y.; Chu, W.; Zhu, J.; Lu, T.; Tai, Y.; and Wang, C. 2022. Seedformer: Patch seeds based point cloud completion with upsample transformer. In European conference on computer vision, 416--432. Springer

Pith tools

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