Pith. sign in

REVIEW 3 major objections 6 minor 3 cited by

Point-GN: A Non-Parametric Network Using Gaussian Positional Encoding for Point Cloud Classification

T0 review · 3 major / 6 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read A fully non-parametric pipeline with Gaussian positional encoding classifies 3D point clouds at 85.29% on ModelNet40 and 85.89% on ScanObjectNN, with zero learnable parameters.

desk verdict A small, clearly-presented tweak to Point-NN, but the headline gains are undermined by hyperparameter tuning on the test sets. read the letter →

arxiv 2412.03056 v2 pith:Q7ILNCPB submitted 2024-12-04 cs.CV cs.AIcs.LGcs.RO

Marzieh Mohammadi , Amir Salarpour This is my paper · ORCID
classification cs.CVcs.AIcs.LGcs.RO
keywords pointcloudclassificationnon-parametricnetworkGaussianpositionalencodingfarthestsamplingk-nearestneighborszerolearnableparametersModelNet40ScanObjectNN
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper seeks to show that 3D point cloud classification does not need learnable parameters at all. It builds a fixed pipeline, Point-GN, that encodes each point with a Gaussian kernel against preset reference coordinates, groups points with farthest-point sampling and k-nearest neighbors, and pools features through four hierarchical stages before a similarity-based classifier compares test features to stored training features. On the standard benchmarks the authors report 85.29% accuracy on ModelNet40 and an average 85.89% on the three ScanObjectNN splits, beating the previous training-free method by 3.5 and up to 21.5 points respectively and coming within about 2 points of the best fully trained model. If true, the result matters because it would mean high-accuracy point cloud recognition is achievable with zero training, suiting real-time and resource-limited devices.

What carries the argument

The load-bearing component is the Gaussian Positional Encoding (GPE), a fixed nonlinear map that turns raw 3D coordinates into a higher-dimensional feature vector by evaluating a Gaussian kernel centered at $V$ uniformly spaced reference points per axis, with width controlled by $\sigma$. This encoding is applied both to the initial points and, after farthest-point sampling and k-nearest-neighbor grouping, to the gathered neighbor coordinates; the encoded neighbor features are combined with the retrieved features via $\Gamma_j \leftarrow \Gamma_j + \gamma(P_j)\odot\gamma(P_j)$, then pooled by mean plus max across neighbors. Four such stages are stacked, each halving the point count, and the final global feature concatenates the per-stage mean-plus-max pooled vectors. Classification is performed without learning: test features are compared by dot product against stored training features, and labels are weighted by $\exp(-\gamma(1-\mathrm{Sim}))$ before a softmax, exactly in the style of Tip-Adapter. Every step is fixed, so the entire network has zero trainable parameters.

What would settle it

Select all four hyperparameters using only a held-out validation split (or cross-validation on training data), freeze the configuration, run Point-GN once on the ModelNet40 and ScanObjectNN test splits, and compare to the reported numbers and to Point-NN under the same protocol. If the accuracies drop materially below 85.29% and 85.89%, or the margin over Point-NN collapses, the central claim is refuted.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central discovery is that replacing the usual sinusoidal positional encoding with a Gaussian positional encoding (GPE)—where each coordinate is mapped through $\exp(-\lVert x_i - v_j\rVert^2/2\sigma^2)$ for $V$ fixed reference points per axis—makes a fully non-parametric point cloud classifier accurate enough to rival trained networks. Combined with farthest point sampling, k-nearest-neighbor grouping, neighbor-mean and neighbor-max pooling, and a training-set similarity classifier with $\exp(-\gamma(1-\mathrm{Sim}))$ activation, the encoder reaches 85.29% on ModelNet40 and average 85.89% on ScanObjectNN across its three splits, improving on Point-NN by 3.5 points on the synthetic set and by up to 21.5 points on the hardest real-world split. The paper also reports that in few-shot settings Point-GN exceeds or matches Point-NN in three of four configurations and beats trained parametric baselines, all with zero learnable parameters.

Load-bearing premise

The reported accuracies are only unbiased if the hyperparameters (neighbor count $K$, GPE dimension, number of stages, and $\sigma$) were chosen without looking at the test sets; the ablation study plots test accuracy as the selection criterion, and if that choice reused the same test data, the 85.29% and 85.89% figures are not an honest measure of generalization.

Editorial extensions

If this is right

  • With zero learned weights, point cloud classifiers can come within about two points of the best trained models on real-world scans and beat the previous training-free approach by 21.5 points on the hardest ScanObjectNN split.
  • Deployment becomes feasible on devices without training infrastructure: inference runs at 301 samples per second on a high-end GPU, and no gradient computation or weight storage is needed.
  • Few-shot classification on ModelNet40 becomes nearly as strong as the non-parametric baseline, and both outperform trained parametric models, suggesting that fixed geometric features carry much of the information needed for small data regimes.
  • A single fixed configuration works across all three ScanObjectNN splits, removing the need for per-dataset tuning or retraining.
  • The Gaussian encoding outperforms sinusoidal encoding in this non-parametric setting, indicating that the choice of positional encoding function matters even without learning.

Reading between the lines

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

  • The same Gaussian positional encoding could be dropped into the input stage of learned 3D networks as a fixed, parameter-free augmentation, potentially improving small-data accuracy or accelerating convergence.
  • Memory of the similarity classifier grows linearly with the number of stored training features, so the method is most naturally suited to few-shot or prototype-based settings; scaling to very large training sets would require a nearest-centroid or hashing approximation.
  • The per-dataset optimal $\sigma$ suggests a multi-scale GPE (several fixed $\sigma$ values concatenated) might improve robustness across object scales without adding trainable parameters.
  • Because the entire pipeline is fixed and hand-designed, the accuracy gap to trained models is a measurable lower bound on what learned features contribute; closing that gap with learned layers could quantify the value of each component.
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 / 6 minor

Summary. This paper proposes Point-GN, a non-parametric point cloud classification pipeline built from Farthest Point Sampling, k-Nearest Neighbors, Gaussian Positional Encoding, and a similarity-based classifier with no learnable parameters. The authors report 85.29% accuracy on ModelNet40 and an average of 85.89% across the three ScanObjectNN splits, claim improvements over the training-free Point-NN baseline, and additionally report few-shot classification results and inference speed comparisons. The central claim is that Gaussian positional encoding, combined with hand-designed feature aggregation, yields a training-free classifier that approaches fully trained models.

Significance. If the reported accuracies were unbiased, the contribution would be notable: a zero-parameter pipeline that comes within about two points of trained models on ScanObjectNN and improves substantially over the prior Point-NN baseline would demonstrate a useful inductive bias for point cloud classification. The paper also includes an ablation study and a public code link. However, the significance is not currently established because the hyperparameters are selected using test-set accuracy in Section 4.7, and at least one headline comparison is overstated relative to the paper's own table. The claimed margins should not be treated as externally validated results until the experimental protocol is corrected.

major comments (3)
  1. [Section 4.7, Figure 5] The ablation study reports test-set accuracy for each choice of K, GPE dimension, number of stages, and sigma, and the text states that for ScanObjectNN the authors "adopted a single configuration for all three splits and aimed to find the best average performance." The same test sets are then used to produce Tables 1 and 2. Because Point-GN has zero learnable parameters, hyperparameter selection is the only form of adaptation, so the headline accuracies (85.29% and 85.89%) and the margins over Point-NN are test-selected rather than unbiased estimates. Please select hyperparameters on a validation split, freeze the configuration, and only then report accuracy on the untouched test sets.
  2. [Table 2 caption and Section 4.4] The caption claims that Point-GN "outperforms all others," but the table lists PointNeXt-S at 87.7% and PointMetaBase-S at 87.9% on PB-T50-RS, both above Point-GN's 86.4%. The caption therefore contradicts the data in the same table. The sentence in Section 4.4 saying Point-GN "consistently outperforms most" is accurate, but the overclaim in the caption and the surrounding discussion of the gap to the best models must be corrected.
  3. [Tables 1 and 2, Sections 4.3-4.4] The Point-NN baseline values (81.8% on ModelNet40 and 71.1/74.9/64.9 on ScanObjectNN) appear lower than the values reported in the original Point-NN paper for the same benchmarks. Since the claimed +3.5 and +21.5 percentage-point improvements are computed against these numbers, the authors must either reproduce Point-NN with their own code under identical conditions and report the protocol, or reconcile the discrepancy with the original publication. Without this, the relative improvement claims are not interpretable.
minor comments (6)
  1. [Section 3.2, Eqs. (4)-(7)] The reference points v_j are used but never defined explicitly as vectors or scalars, and the concatenation notation [gamma_x, gamma_y, gamma_z]_{j=1}^V is ambiguous. Please define v_j in R^3 and write the concatenation out explicitly.
  2. [Section 3.3.3, Eq. (12)] The symbol gamma(P_j) is used both for the retrieved features from Eq. (11) and for a new Gaussian encoding of the retrieved coordinates, which makes the update rule ambiguous. Please use distinct notation for these two quantities.
  3. [Section 3.3.1] The text says reference points are "strategically chosen or learned by the model," but the model has no learnable parameters. This should be restated to say they are hand-set, since the claim of zero learnable parameters is central to the paper.
  4. [Section 4.7, Figure 5(b)] The x-axis of Figure 5(b) is labeled only "Dimension" without stating whether the values are V or V times 3. The text discusses a dimension of 27, so the axis definition should be clarified.
  5. [Table 3] The table reports mean accuracy across 10 runs, but no standard deviations are given. Please add standard deviations, especially for the 5-way 20-shot configuration where Point-GN and Point-NN are reported as tied.
  6. [Figure 4] The inference speed comparison is shown only as a bar plot without numeric labels or error bars. A table with the measured samples per second for each dataset would make the claimed speed advantage reproducible and easier to verify.

Circularity Check

1 steps flagged · score 6.0 of 10

Headline accuracies are selected on the test sets in Section 4.7, making the claimed gains over Point-NN fitted values rather than unbiased predictions.

  1. fitted input called prediction [Section 4.7 (Ablation Study, Fig. 5) feeding Section 4.4 and Tables 1-2]
    "Figure 5. Ablation study results showing the sensitivity of Point-GN's performance to key hyperparameters: (a) Number of neighbors (K), (b) Dimension of Gaussian Positional Encoding (GPE), (c) Number of stages, and (d) Sigma (σ). We compare the performance of the model on ModelNet40 [34] (orange) and ScanObjectNN [28] (cyan) datasets, showing both the average and best performances. ... The 4-stage configuration achieves the highest accuracy of 85.3% on the ModelNet40 dataset and 85.9% on the ScanObjectNN dataset ..."

    Point-GN has zero learnable parameters, so K, GPE dimension, stage count, and sigma are the only adaptive degrees of freedom. Section 4.7 tunes these by reading 'Accuracy (%)' on the ModelNet40 and ScanObjectNN test sets (Fig. 5), and Section 4.4 then fixes 'a single configuration' giving 'the best average performance.' That same configuration is used to report the headline 85.29%/85.89% accuracies in the abstract and Tables 1-2. The reported numbers are thus the maximands of the selection procedure evaluated on the same test sets, not out-of-sample measurements. With no validation split or held-out data, the +3.5% ModelNet40 and +21.5% PB-T50-RS margins over Point-NN are in-sample fits by construction, so the central quantitative claim reduces to the tuning rule.

full rationale

The derivation of Point-GN itself (Eqs. 4-14 for GPE/FPS/KNN pooling and Eqs. 17-19 for the similarity classifier) is self-contained and does not encode the target accuracies; those equations are not circular. The circularity is in the evaluation protocol: since the network has no trainable weights, hyperparameter selection is the entire learning process, and the paper performs that selection on the same test sets whose final accuracy it reports. This matches the 'fitted input called prediction' pattern rather than self-citation: Point-NN [43] and Tip-Adapter [41] are external baselines/adaptations and are not used to justify the headline numbers. If the hyperparameters had been fixed on a validation split or pre-registered before testing, the claimed gains would be independently checkable. One additional non-circular consistency issue: Table 2's caption says Point-GN 'outperforms all others' although PointNeXt-S scores 87.7 and PointMetaBase-S scores 87.9 on PB-T50-RS versus 86.4 for Point-GN; this does not affect the circularity verdict. Score 6 is appropriate: the headline empirical claims partially reduce, by construction, to the test-set-based selection rule.

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

Point-GN has no learned weights, but its performance depends on at least six hand-set or test-tuned quantities: sigma, K, GPE dimension, stage count, reference point placement, and classifier gamma. Most importantly, four of these were selected by maximizing accuracy on the test sets used for the final tables, so the evaluation supplies information the method is supposed to predict. The axioms list the assumptions behind the GPE feature map and the similarity classifier.

free parameters (6)
  • Gaussian kernel width sigma = 0.35 or 0.4 (ModelNet40); 0.3 (ScanObjectNN)
    Set by test-accuracy ablation, Section 4.7, Fig. 5d.
  • Number of neighbors K = 120
    Set by test-accuracy ablation, Section 4.7, Fig. 5a.
  • GPE dimension V = 27
    Set by test-accuracy ablation, Section 4.7, Fig. 5b.
  • Number of stages = 4
    Set by test-accuracy ablation, Section 4.7, Fig. 5c.
  • Reference points v_j = uniform in [-1, 1] (stated as 'often')
    Hand-chosen centers for the Gaussian kernel, Section 3.3.1; no selection rule or justification is given.
  • Classifier scaling gamma = not reported
    Appears in Eq. 18; no value or tuning procedure is provided.
assumptions (4)
  • ad hoc to paper Gaussian RBF encoding of coordinates with hand-set centers and width produces features that separate classes better than sinusoidal encoding in a non-parametric pipeline.
    No derivation or independent evidence; this is the core assumption behind GPE, introduced in Section 3.2 and validated only through test-set ablations.
  • domain assumption Cosine similarity to stored training features is a valid classifier without calibration.
    Adopted from Point-NN [43] and Tip-Adapter [41] in Section 3.4; no analysis of bias, calibration, or sensitivity to feature scale is given.
  • ad hoc to paper Test-set accuracy is a valid model selection criterion.
    The paper selects all hyperparameters by test accuracy in Section 4.7; this violates standard evaluation practice and inflates reported accuracy.
  • domain assumption FPS and KNN on normalized coordinates preserve class-discriminative geometry.
    Standard in point cloud pipelines, used in Section 3.3.2; accepted without independent verification.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Point-GN: A Non-Parametric Network Using Gaussian Positional Encoding for Point Cloud Classification." pith.science (2026). https://pith.science/paper/Q7ILNCPB

@misc{pith2026241203056,
  author       = {Pith},
  title        = {Pith review of: Point-GN: A Non-Parametric Network Using Gaussian Positional Encoding for Point Cloud Classification},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/Q7ILNCPB}},
  note         = {Machine review of arXiv:2412.03056}
}
read the original abstract

This paper introduces Point-GN, a novel non-parametric network for efficient and accurate 3D point cloud classification. Unlike conventional deep learning models that rely on a large number of trainable parameters, Point-GN leverages non-learnable components-specifically, Farthest Point Sampling (FPS), k-Nearest Neighbors (k-NN), and Gaussian Positional Encoding (GPE)-to extract both local and global geometric features. This design eliminates the need for additional training while maintaining high performance, making Point-GN particularly suited for real-time, resource-constrained applications. We evaluate Point-GN on two benchmark datasets, ModelNet40 and ScanObjectNN, achieving classification accuracies of 85.29% and 85.89%, respectively, while significantly reducing computational complexity. Point-GN outperforms existing non-parametric methods and matches the performance of fully trained models, all with zero learnable parameters. Our results demonstrate that Point-GN is a promising solution for 3D point cloud classification in practical, real-time environments.

Figures

Figures reproduced from arXiv: 2412.03056 by the authors.

Figure 1
Figure 1. Illustration of the proposed non-parametric network for [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Overview of the architecture of our Non-Parametric Fea [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Illustration of the non-parametric classifier pipeline. The [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Test Speed (samples per second) on ScanObjectNN [28] and ModelNet40 [34] datasets. The plot compares the in￾ference speed of Point-NN [43] and Point-GN on four different datasets. Point-GN shows significant improvements in inference speed across all datasets. Point-GN,…
Figure 5
Figure 5. Figure 5: Ablation study results showing the sensitivity of Point-GN’s performance to key hyperparameters: (a) Number of neighbors ( [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 3 Pith papers

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

  1. Synthetic LiDAR Data Generation and Deterministic Downsampling for Point Cloud Classification on the Edge

    cs.LG 2026-08 conditional novelty 4.0 of 10

    The authors combine a BLAINDER-based synthetic LiDAR variant of ModelNet40 with a standalone pretrained Critical Point Layer frontend, reporting 88.36% accuracy and about 50 FPS on a Raspberry Pi 5.

  2. Enhancing 3D Point Cloud Classification with ModelNet-R and Point-SkipNet

    cs.CV 2025-09 reject novelty 4.0 of 10

    ModelNet-R cleans five ModelNet40 classes and Point-SkipNet is a lightweight point cloud classifier, but the reported gains over ModelNet are not proven to reflect data quality rather than a changed test set.

  3. Point-LN: A Lightweight Framework for Efficient Point Cloud Classification Using Non-Parametric Positional Encoding

    cs.CV 2025-01 conditional novelty 3.0 of 10

    A hybrid point cloud classifier with non-learnable positional encodings and a tiny learnable classifier reaches competitive accuracy with 0.8M parameters.

Reference graph

Works this paper leans on

47 extracted references · 34 canonical work pages · cited by 3 Pith papers

  1. [1]

    Mohammed Aldeen, Pedram MohajerAnsari, Jin Ma, Mashrur Chowdhury, Long Cheng, and Mert D. Pes ´e. An initial exploration of employing large multimodal models in defending against autonomous vehicles attacks. In 2024 IEEE Intelligent Vehicles Symposium (IV), pages 3334–3341,

    work page 2024

  2. [2]

    3dmfv: Three-dimensional point cloud classification in real-time using convolutional neural networks

    Yizhak Ben-Shabat, Michael Lindenbaum, and Anath Fis- cher. 3dmfv: Three-dimensional point cloud classification in real-time using convolutional neural networks. IEEE Robotics and Automation Letters, 3(4):3145–3152, 2018. 6

    work page 2018

  3. [3]

    Gvcnn: Group-view convolutional neural networks for 3d shape recognition

    Yifan Feng, Zizhao Zhang, Xibin Zhao, Rongrong Ji, and Yue Gao. Gvcnn: Group-view convolutional neural networks for 3d shape recognition. In Proceedings of the IEEE con- ference on computer vision and pattern recognition , pages 264–272, 2018. 2

    work page 2018

  4. [4]

    Revisiting point cloud shape classification with a simple and effective baseline

    Ankit Goyal, Hei Law, Bowei Liu, Alejandro Newell, and Jia Deng. Revisiting point cloud shape classification with a simple and effective baseline. In International Conference on Machine Learning, pages 3809–3820. PMLR, 2021. 2

    work page 2021

  5. [5]

    A deep learning network for point cloud of medicine structure

    Jia Guo, Xuanxia Yao, Mengyu Shen, Jiafei Wang, and Wanyou Liao. A deep learning network for point cloud of medicine structure. In 2018 9th International Conference on Information Technology in Medicine and Education (ITME), pages 683–687. IEEE, 2018. 1

    work page 2018

  6. [6]

    Randla-net: Efficient semantic segmentation of large-scale point clouds

    Qingyong Hu, Bo Yang, Linhai Xie, Stefano Rosa, Yulan Guo, Zhihua Wang, Niki Trigoni, and Andrew Markham. Randla-net: Efficient semantic segmentation of large-scale point clouds. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages 11108– 11117, 2020. 2

    work page 2020

  7. [7]

    Pointcnn: Convolution on x-transformed points

    Yangyan Li, Rui Bu, Mingchao Sun, Wei Wu, Xinhan Di, and Baoquan Chen. Pointcnn: Convolution on x-transformed points. Advances in neural information processing systems, 31, 2018. 2, 6, 7

    work page 2018

  8. [8]

    Meta architecture for point cloud analysis

    Haojia Lin, Xiawu Zheng, Lijiang Li, Fei Chao, Shanshan Wang, Yan Wang, Yonghong Tian, and Rongrong Ji. Meta architecture for point cloud analysis. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 17682–17691, 2023. 6, 7

    work page 2023

Show all 47 references
  1. [9]

    Point- voxel cnn for efficient 3d deep learning

    Zhijian Liu, Haotian Tang, Yujun Lin, and Song Han. Point- voxel cnn for efficient 3d deep learning. Advances in neural information processing systems, 32, 2019. 2

    2019

  2. [10]

    Point- ngcnn: Deep convolutional networks on 3d point clouds with neighborhood graph filters

    Qiang Lu, Chao Chen, Wenjun Xie, and Yuetong Luo. Point- ngcnn: Deep convolutional networks on 3d point clouds with neighborhood graph filters. Computers & Graphics, 86:42– 51, 2020. 2

    2020

  3. [11]

    Rethinking network design and local geometry in point cloud: A simple residual mlp framework

    Xu Ma, Can Qin, Haoxuan You, Haoxi Ran, and Yun Fu. Rethinking network design and local geometry in point cloud: A simple residual mlp framework. arXiv preprint arXiv:2202.07123, 2022. 1, 6, 7

    2022 arXiv

  4. [12]

    V oxnet: A 3d con- volutional neural network for real-time object recognition

    Daniel Maturana and Sebastian Scherer. V oxnet: A 3d con- volutional neural network for real-time object recognition. In 2015 IEEE/RSJ international conference on intelligent robots and systems (IROS), pages 922–928. IEEE, 2015. 2

    2015

  5. [13]

    Nerf: Representing scenes as neural radiance fields for view syn- thesis

    Ben Mildenhall, Pratul P Srinivasan, Matthew Tancik, Jonathan T Barron, Ravi Ramamoorthi, and Ren Ng. Nerf: Representing scenes as neural radiance fields for view syn- thesis. Communications of the ACM , 65(1):99–106, 2021. 2

    2021

  6. [14]

    Discovering new shadow patterns for black-box attacks on lane detection of autonomous vehicles

    Pedram MohajerAnsari, Alkim Domeke, Jan de V oor, Arka- jyoti Mitra, Grace Johnson, Amir Salarpour, Habeeb Olu- fowobi, Mohammad Hamad, and Mert D Pes ´e. Discovering new shadow patterns for black-box attacks on lane detection of autonomous vehicles. arXiv preprint arXiv:2409.18248,

  7. [15]

    Curvenet: Curvature-based multitask learning deep networks for 3d object recognition

    AAM Muzahid, Wanggen Wan, Ferdous Sohel, Lianyao Wu, and Li Hou. Curvenet: Curvature-based multitask learning deep networks for 3d object recognition. IEEE/CAA Journal of Automatica Sinica, 8(6):1177–1187, 2020. 1

    2020

  8. [16]

    Deep neural networks are easily fooled: High confidence predictions for unrecognizable images

    Anh Nguyen, Jason Yosinski, and Jeff Clune. Deep neural networks are easily fooled: High confidence predictions for unrecognizable images. In Proceedings of the IEEE con- ference on computer vision and pattern recognition , pages 427–436, 2015. 2

    2015

  9. [17]

    Differentiable volumetric rendering: Learn- ing implicit 3d representations without 3d supervision

    Michael Niemeyer, Lars Mescheder, Michael Oechsle, and Andreas Geiger. Differentiable volumetric rendering: Learn- ing implicit 3d representations without 3d supervision. In Proceedings of the IEEE/CVF conference on computer vi- sion and pattern recognition, pages 3504–3515, 2020. 2

    2020

  10. [18]

    Point-planenet: Plane kernel based convolutional neural network for point clouds analysis

    SM Moein Peyghambarzadeh, Fatemeh Azizmalayeri, Has- san Khotanlou, and Amir Salarpour. Point-planenet: Plane kernel based convolutional neural network for point clouds analysis. Digital Signal Processing, 98:102633, 2020. 2

    2020

  11. [19]

    Frustum pointnets for 3d object detection from rgb- d data

    Charles R Qi, Wei Liu, Chenxia Wu, Hao Su, and Leonidas J Guibas. Frustum pointnets for 3d object detection from rgb- d data. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 918–927, 2018. 1

    2018

  12. [20]

    Pointnet: Deep learning on point sets for 3d classification and segmentation

    Charles R Qi, Hao Su, Kaichun Mo, and Leonidas J Guibas. Pointnet: Deep learning on point sets for 3d classification and segmentation. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 652–660,

  13. [21]

    Pointnet++: Deep hierarchical feature learning on point sets in a metric space

    Charles Ruizhongtai Qi, Li Yi, Hao Su, and Leonidas J Guibas. Pointnet++: Deep hierarchical feature learning on point sets in a metric space. Advances in neural information processing systems, 30, 2017. 1, 2, 6, 7

    2017

  14. [22]

    Pointnext: Revisiting pointnet++ with improved training and scaling strategies

    Guocheng Qian, Yuchen Li, Houwen Peng, Jinjie Mai, Hasan Hammoud, Mohamed Elhoseiny, and Bernard Ghanem. Pointnext: Revisiting pointnet++ with improved training and scaling strategies. Advances in neural informa- tion processing systems, 35:23192–23204, 2022. 6

    2022

  15. [23]

    Geometric back- projection network for point cloud classification

    Shi Qiu, Saeed Anwar, and Nick Barnes. Geometric back- projection network for point cloud classification. IEEE Transactions on Multimedia, 24:1943–1955, 2021. 6

    1943

  16. [24]

    Long-term estimation of human spatial interactions through multiple laser ranging sensors

    Amir Salarpour, Hassan Khotanlou, and Nikolaos Mavridis. Long-term estimation of human spatial interactions through multiple laser ranging sensors. In2014 International Confer- ence on Robotics and Emerging Allied Technologies in Engi- neering (iCREATE), pages 109–114. IEEE, 2014. 2

    2014

  17. [25]

    Self-supervised few-shot learning on point clouds

    Charu Sharma and Manohar Kaul. Self-supervised few-shot learning on point clouds. Advances in Neural Information Processing Systems, 33:7212–7221, 2020. 7

    2020

  18. [26]

    Multi-view convolutional neural networks 9 for 3d shape recognition

    Hang Su, Subhransu Maji, Evangelos Kalogerakis, and Erik Learned-Miller. Multi-view convolutional neural networks 9 for 3d shape recognition. In Proceedings of the IEEE in- ternational conference on computer vision , pages 945–953,

  19. [27]

    Kpconv: Flexible and deformable convolution for point clouds

    Hugues Thomas, Charles R Qi, Jean-Emmanuel Deschaud, Beatriz Marcotegui, Franc ¸ois Goulette, and Leonidas J Guibas. Kpconv: Flexible and deformable convolution for point clouds. In Proceedings of the IEEE/CVF international conference on computer vision, pages 6411–6420, 2019. 2

    2019

  20. [28]

    Revisiting point cloud classification: A new benchmark dataset and classification model on real-world data

    Mikaela Angelina Uy, Quang-Hieu Pham, Binh-Son Hua, Thanh Nguyen, and Sai-Kit Yeung. Revisiting point cloud classification: A new benchmark dataset and classification model on real-world data. In Proceedings of the IEEE/CVF international conference on computer vision , pages 1...

    2019

  21. [29]

    Shape completion enabled robotic grasping

    Jacob Varley, Chad DeChant, Adam Richardson, Joaqu ´ın Ruales, and Peter Allen. Shape completion enabled robotic grasping. In 2017 IEEE/RSJ international conference on intelligent robots and systems (IROS) , pages 2442–2447. IEEE, 2017. 1

    2017

  22. [30]

    Attention is all you need

    A Vaswani. Attention is all you need. Advances in Neural Information Processing Systems, 2017. 2

    2017

  23. [31]

    Msnet: Multi-scale convolutional network for point cloud classifica- tion

    Lei Wang, Yuchun Huang, Jie Shan, and Liu He. Msnet: Multi-scale convolutional network for point cloud classifica- tion. Remote Sensing, 10(4):612, 2018. 2

    2018

  24. [32]

    Dynamic graph cnn for learning on point clouds

    Yue Wang, Yongbin Sun, Ziwei Liu, Sanjay E Sarma, Michael M Bronstein, and Justin M Solomon. Dynamic graph cnn for learning on point clouds. ACM Transactions on Graphics (tog), 38(5):1–12, 2019. 2, 6, 7

    2019

  25. [33]

    Learning a probabilistic latent space of object shapes via 3d generative-adversarial modeling

    Jiajun Wu, Chengkai Zhang, Tianfan Xue, Bill Freeman, and Josh Tenenbaum. Learning a probabilistic latent space of object shapes via 3d generative-adversarial modeling. Ad- vances in neural information processing systems , 29, 2016. 7

    2016

  26. [34]

    3d shapenets: A deep representation for volumetric shapes

    Zhirong Wu, Shuran Song, Aditya Khosla, Fisher Yu, Lin- guang Zhang, Xiaoou Tang, and Jianxiong Xiao. 3d shapenets: A deep representation for volumetric shapes. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 1912–1920, 2015. 2, 5, 6, 7, 8

    1912

  27. [35]

    Walk in the cloud: Learning curves for point clouds shape analysis

    Tiange Xiang, Chaoyi Zhang, Yang Song, Jianhui Yu, and Weidong Cai. Walk in the cloud: Learning curves for point clouds shape analysis. In Proceedings of the IEEE/CVF in- ternational conference on computer vision , pages 915–924,

  28. [36]

    A re- view: Point cloud-based 3d human joints estimation

    Tianxu Xu, Dong An, Yuetong Jia, and Yang Yue. A re- view: Point cloud-based 3d human joints estimation. Sen- sors, 21(5):1684, 2021. 1

    2021

  29. [37]

    Modeling point clouds with self-attention and gumbel subset sampling

    Jiancheng Yang, Qiang Zhang, Bingbing Ni, Linguo Li, Jinxian Liu, Mengdie Zhou, and Qi Tian. Modeling point clouds with self-attention and gumbel subset sampling. In Proceedings of the IEEE/CVF conference on computer vi- sion and pattern recognition, pages 3323–3332, 2019. 2

    2019

  30. [38]

    Fold- ingnet: Point cloud auto-encoder via deep grid deformation

    Yaoqing Yang, Chen Feng, Yiru Shen, and Dong Tian. Fold- ingnet: Point cloud auto-encoder via deep grid deformation. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 206–215, 2018. 7

    2018

  31. [39]

    3d medical point transformer: Introducing convolution to attention networks for medical point cloud analysis

    Jianhui Yu, Chaoyi Zhang, Heng Wang, Dingxin Zhang, Yang Song, Tiange Xiang, Dongnan Liu, and Weidong Cai. 3d medical point transformer: Introducing convolution to attention networks for medical point cloud analysis. arXiv preprint arXiv:2112.04863, 2021. 1

    2021 arXiv

  32. [40]

    Point-bert: Pre-training 3d point cloud transformers with masked point modeling

    Xumin Yu, Lulu Tang, Yongming Rao, Tiejun Huang, Jie Zhou, and Jiwen Lu. Point-bert: Pre-training 3d point cloud transformers with masked point modeling. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 19313–19322, 2022. 2

    2022

  33. [41]

    Tip-adapter: Training-free clip-adapter for better vision- language modeling

    Renrui Zhang, Rongyao Fang, Wei Zhang, Peng Gao, Kunchang Li, Jifeng Dai, Yu Qiao, and Hongsheng Li. Tip-adapter: Training-free clip-adapter for better vision- language modeling. arXiv preprint arXiv:2111.03930, 2021. 5

    2021 arXiv

  34. [42]

    Pointclip: Point cloud understanding by clip

    Renrui Zhang, Ziyu Guo, Wei Zhang, Kunchang Li, Xu- peng Miao, Bin Cui, Yu Qiao, Peng Gao, and Hongsheng Li. Pointclip: Point cloud understanding by clip. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 8552–8562, 2022. 2

    2022

  35. [43]

    Parameter is not all you need: Starting from non-parametric networks for 3d point cloud analysis

    Renrui Zhang, Liuhui Wang, Ziyu Guo, Yali Wang, Peng Gao, Hongsheng Li, and Jianbo Shi. Parameter is not all you need: Starting from non-parametric networks for 3d point cloud analysis. arXiv preprint arXiv:2303.08134, 2023. 2, 5, 6, 7

    2023 arXiv

  36. [44]

    Shellnet: Efficient point cloud convolutional neural networks using concentric shells statistics

    Zhiyuan Zhang, Binh-Son Hua, and Sai-Kit Yeung. Shellnet: Efficient point cloud convolutional neural networks using concentric shells statistics. In Proceedings of the IEEE/CVF international conference on computer vision , pages 1607– 1616, 2019. 2

    2019

  37. [45]

    A camera and lidar data fusion method for railway ob- ject detection

    Wang Zhangyu, Yu Guizhen, Wu Xinkai, Li Haoran, and Li Da. A camera and lidar data fusion method for railway ob- ject detection. IEEE Sensors Journal, 21(12):13442–13454,

  38. [46]

    Point transformer

    Hengshuang Zhao, Li Jiang, Jiaya Jia, Philip HS Torr, and Vladlen Koltun. Point transformer. In Proceedings of the IEEE/CVF international conference on computer vision, pages 16259–16268, 2021. 2

    2021

  39. [47]

    V oxelnet: End-to-end learning for point cloud based 3d object detection

    Yin Zhou and Oncel Tuzel. V oxelnet: End-to-end learning for point cloud based 3d object detection. In Proceedings of the IEEE conference on computer vision and pattern recog- nition, pages 4490–4499, 2018. 1 10

    2018

Pith tools

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