Pith. sign in

REVIEW 4 major objections 5 minor 45 references

UPP: Unified Point-Level Prompting for Robust Point Cloud Analysis

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

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

Pith's one-line read The paper claims that reformulating point cloud denoising and completion as point-level prompts lets a frozen pre-trained backbone beat full fine-tuning on noisy and incomplete classification with only 1.4M trainable parameters.

desk verdict UPP is a genuinely new point-level prompting architecture that reports consistent gains on noisy/incomplete point cloud classification, but the headline comparison gives it up to 50 extra training epochs plus dense geometric supervision the baselines don't get, so the efficiency claim isn't isolated. read the letter →

arxiv 2507.18997 v1 pith:47NK7QH5 submitted 2025-07-25 cs.CV

Zixiang Ai , Zhenyu Cui , Yuxin Peng , Jiahuan Zhou This is my paper · ORCID
classification cs.CV
keywords pointcloudanalysisparameter-efficientfine-tuningpoint-levelpromptingdenoisingcompletion3DclassificationnoisycloudsShape-AwareAttention
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

Pre-trained point cloud models lose accuracy when input scans are noisy or missing points, a common situation with real 3D sensors. The paper proposes UPP, a parameter-efficient fine-tuning method that treats denoising and completion as point-level prompts rather than as separate pre-processing models. A Rectification Prompter predicts per-point displacement vectors and masks unreliable points, a Completion Prompter generates missing points using a repurposed pre-trained decoder, and a Shape-Aware Unit adapts the frozen backbone's features through spatial-distance attention. With only 1.4 million trainable parameters, UPP reports higher classification accuracy than full fine-tuning on Noisy ModelNet40, Noisy ShapeNet55, and ScanObjectNN, across three different pre-trained backbones. The paper also shows that applying rectification before completion, and the addition of the Shape-Aware Unit, account for most of the gain.

What carries the argument

The machinery is three lightweight modules inserted into a frozen pre-trained point cloud transformer: the Rectification Prompter, which estimates per-point displacement vectors $v_r \in \mathbb{R}^3$ and masks points with $\|v_r\| > \tau$; the Completion Prompter, which reuses the discarded decoder of the masked autoencoder to produce coarse centers and dense point prompts $x_m$; and the Shape-Aware Unit, which adds trainable prompt tokens $k_i$ to each block and combines feature-similarity attention with spatial-distance K-nearest-neighbor interpolation. The completion losses are Chamfer distances between predicted and ground-truth clouds, the rectification loss is an L2 distance to the nearest clean-surface points, and the total loss is the sum of the two plus the downstream cross-entropy. This design lets the model move and add points in the input space while keeping the backbone weights frozen.

What would settle it

Run the Table 1 comparison with identical total epochs for every method and the same auxiliary geometric losses; if UPP's Noisy ModelNet40 accuracy drops to or below the 89.42% full-fine-tuning baseline, the advertised gain comes from the extra training budget, not from point-level prompting.

Watch

Extended reading notes

Core claim

UPP's central claim is that denoising and completion should not be separate pre-processing stages: they can be expressed as point-level prompts inside the analysis model, so the enhancement and the downstream task share one frozen backbone. Given a noisy, incomplete point cloud, the Rectification Prompter predicts a per-point rectification vector, keeps only points whose predicted displacement magnitude is below a threshold, and moves those points toward the estimated clean surface. The Completion Prompter then repurposes the pre-trained masked-autoencoder decoder to predict coarse missing centers and dense auxiliary points, and the fused point set is resampled with farthest point sampling. The Shape-Aware Unit prepends prompt tokens in every transformer block and augments feature-similarity attention with spatial-distance K-nearest-neighbor attention, which the paper argues is naturally insensitive to outliers. On Noisy ModelNet40 with the Point-MAE backbone this reaches 92.95% accuracy with 1.4M trainable parameters, versus 89.42% for full fine-tuning and 86.43% for DAPT.

Load-bearing premise

The comparison assumes a matched training budget, but the staged optimization gives UPP up to 50 extra epochs plus auxiliary Chamfer supervision that the baselines do not receive, so the accuracy margin may partly reflect extra training rather than the prompting mechanism alone.

Editorial extensions

If this is right

  • If UPP's results hold, freezing the backbone while injecting point-level prompts is enough to exceed full fine-tuning on noisy and incomplete classification, so full fine-tuning is not necessary for robustness on these benchmarks.
  • The method's 1.4M trainable parameters, a reduction of more than 95% versus full fine-tuning, imply that the accuracy gain does not require task-specific denoising and completion networks with their own feature extractors.
  • The ordering result, where rectification first reaches 92.95% versus 91.18% for completion first, implies that cleaning input geometry before completing the shape is a design principle for unified enhancement pipelines.
  • On clean data, the Shape-Aware Unit alone reaches 94.2% on ModelNet40 with a Point-FEMAE backbone and 0.6M parameters, suggesting the module is also a general parameter-efficient fine-tuning method, not only a noise-handling device.
  • Part segmentation results on noisy ShapeNetPart show UPP beats other PEFT methods but still trails full fine-tuning in fine-grained tasks, so the claimed advantage is strongest for classification.

Reading between the lines

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

  • A testable extension is to isolate the source of the gain by giving DAPT and full fine-tuning the same auxiliary Chamfer-distance supervision and extra epochs; if the gap narrows, the contribution is partly the auxiliary geometric loss rather than the prompting mechanism itself.
  • The masking of low-reliability points by a hard threshold could be replaced with a soft, differentiable weighting, which might preserve more thin structures such as chair legs or table edges.
  • The same point-level prompting recipe could be applied to other corruption types, such as varying occlusion rates or sensor-specific noise patterns, and to 3D detection, where noisy partial scans are also common.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 5 minor

Summary. The paper proposes UPP, a parameter-efficient fine-tuning (PEFT) method that reformulates point cloud denoising and completion as point-level prompting for downstream classification and segmentation. Three modules are introduced: a Rectification Prompter (Eqs. 1-4) that moves noisy points along predicted vectors, a Completion Prompter (Eqs. 5-9) that reconstructs missing regions using the repurposed MAE decoder, and a Shape-Aware Unit (Eqs. 10-12) that adds spatial-distance-based attention and adapters into each frozen transformer block. Training combines the task loss with rectification and completion losses (Eq. 14). Experiments on synthetic Noisy ModelNet40, Noisy ShapeNet55, real-world ScanObjectNN, and noisy ShapeNetPart report that UPP outperforms full fine-tuning and prior PEFT methods with only 1.4M trainable parameters. The supplementary details the training schedule, additional ablations, and a clean-data PEFT comparison.

Significance. If the reported comparisons are valid, the contribution is a practical one: a single frozen-backbone framework that performs enhancement and analysis jointly, with clearly specified losses and a released codebase. The explicit mathematical formulation of the losses and the careful description of the prompting mechanism are strengths, as is the evaluation across multiple backbone choices and noise regimes. However, the central efficiency-and-accuracy claim is not yet isolated from the training budget and auxiliary supervision the method receives; the margin over full fine-tuning could plausibly be driven by extra epochs and dense geometric losses rather than by the prompting design. The paper is therefore promising but requires a substantially fairer experimental protocol before its headline claims are supported.

major comments (4)
  1. [Supplementary 'Staged Optimization Strategy' and Table 4] The training budget for UPP is not matched to the baselines. The supplementary states: 'We add 50 epochs to optimize the point-level promoters' (20 epochs for both prompters, then 30 for the Completion Prompter), while Table 4 fixes Training epochs at 300 for all methods and the text claims 'identical hyper-parameters and training strategies are applied across fine-tuning and proposed methods.' These statements are internally inconsistent: UPP receives up to 350 epochs in total. Since Table 1 reports margins of +3.53 over full fine-tuning on Noisy ModelNet40 and +2.27 on Noisy ShapeNet55, an additional 50 epochs of training is a plausible alternative explanation for the improvement. This is load-bearing because the paper's central selling point is higher accuracy with 95% fewer trainable parameters; trainable parameter count is not the same as total training budget. Please rerun all baselines under a matched total-epoch budget (e.g., 350 epochs for every method), or restrict UPP to 300 total epochs including the prompter warm-up, and report the resulting numbers.
  2. [Eq. (14) and the auxiliary losses in Eqs. (4) and (9)] The total loss L = Lrect + Lcomp + Ltask gives UPP dense geometric supervision that the baselines do not receive. Lrect uses the clean point cloud as target and a separate noisy-point set, while Lcomp uses the ground-truth complete point cloud Pgt and the missing part Pm; full fine-tuning, DAPT, and Point-PEFT are trained with cross-entropy alone. Because UPP's synthetic corruption during training (24 outliers, 64 surface noise, 25% missing points, and supplementary random cropping of 25%-50%) matches the corruption at test time, the gain could come from the auxiliary losses on the test distribution rather than from the point-level prompting mechanism. Please add an ablation in which the baselines receive analogous auxiliary denoising/completion supervision (or in which UPP is trained with only Ltask) to isolate the contribution of the prompting design.
  3. [Section 5.4 and all experimental tables] No error bars, seeds, or statistical significance tests are reported anywhere in the paper or supplement, and the ablation design in Table 3 and the prompting-order ablation (Table 7) are selected on Noisy ModelNet40, the same benchmark that produces the headline numbers. With component increments of 1.50%, 1.50%, and 1.54% in Table 3 and margins as small as +0.67 in Table 1, single-run differences may be within run-to-run variance. Please report mean and standard deviation over at least three seeds for the main tables and the key ablations, and specify whether the model selection is performed on a validation split or on the test set.
  4. [Supplementary Table 5] On Noisy ShapeNetPart, UPP (82.2/84.4 with Point-MAE and 82.5/84.8 with Point-FEMAE) does not surpass full fine-tuning (83.3/85.6 and 83.5/85.9), and the text acknowledges that PEFT methods including UPP 'exhibit greater susceptibility to noise and incompleteness compared to full fine-tuning.' This is stated in the supplement but not mentioned in the main paper's conclusion, which claims general 'superiority' over existing methods. The main text should either include the segmentation results or explicitly qualify the claim to the classification benchmarks.
minor comments (5)
  1. [Table 2] The Point-FEMAE row is duplicated, and PCP-MAE is labeled with reference [41] although it is reference [43] in the bibliography; please correct the table and the citation indices.
  2. [Table 1] The grouping is confusing: the Point-MAE row listed under 'Full Fine-Tuning (FFT)' and the 'Point-MAE (baseline)' row under 'Parameter-Efficient Fine-Tuning' are identical in the table yet represent different training regimes; please restructure the table so that the baseline configuration is unambiguous.
  3. [Supplementary 'Staged Optimization Strategy'] The phrase 'we optionally enable the training of the two point-level promoters with the Shape-Aware Unit when the learning rate narrows to 0.0001' is vague; please specify exactly which epochs or which LR milestones trigger this phase and whether the 50 extra epochs are included in the 300 reported in Table 4.
  4. [Figure 3] The y-axis range and the numeric labels under each outlier-noise level are difficult to read, and the relationship between the plotted percentages and the 'Outlier Noise Number / Input Resolution' axis label is unclear; please annotate the figure more precisely.
  5. [Section 2.2 and abstract] The abstract and introduction describe the conventional denoising/completion ensemble as 'failing' and 'inefficient,' but the quantitative comparison to such ensemble methods is not shown; adding one such comparison, even in the supplement, would strengthen the positioning relative to the described baseline.

Circularity Check

0 steps flagged · score 0.0 of 10

No circular derivation; the headline comparison is empirical and does not reduce to its inputs by construction.

full rationale

UPP is an empirical method paper rather than a derivation chain. The Rectification Prompter and Completion Prompter are trained with auxiliary geometric losses (Eq. 4 and Eq. 9) on corrupted inputs and clean/missing ground truth, and the frozen-backbone model is then evaluated on held-out test sets; this is standard supervised training, not a fitted quantity being renamed as a prediction. The claimed superiority over full fine-tuning and other PEFT methods rests on reported accuracies against official-code baselines, not on a theorem or an equation that reduces to its own inputs. The self-citations (Refs. [1,2]) appear as background enumerations in Related Work and are not load-bearing for any central claim. The main legitimate concerns are experimental fairness and generalization: the supplementary 'Staged Optimization Strategy' grants UPP up to 50 additional epochs and dense geometric supervision (Lrect + Lcomp) that baselines do not receive, and the prompting order is selected on the Noisy ModelNet40 test set. These are evaluation-design and generalization risks, not circularity, because nothing in the paper's definitions or equations makes the reported accuracy true by construction.

Assumptions & free parameters 7 free parameters · 5 assumptions · 0 invented entities

The method rests on several unverified domain assumptions: frozen pre-trained weights transfer to noisy inputs, the synthetic corruption model approximates real degradation, the k-NN displacement to the clean surface is a valid rectification target, and the pre-trained MAE decoder can be reused for dense completion without fine-tuning. The numerical results also depend on a set of unreported or tuned hyperparameters (alpha, tau, K, rank r, block allocation, number of coarse centers, and the staged epoch schedule), so the reader cannot fully separate the method's contribution from the tuning effort.

free parameters (7)
  • alpha (blending factor in Eq. 3) = not reported
    Controls how far points are moved during rectification; chosen without a report in the text, likely tuned on the validation or test set.
  • tau (rectification threshold in Eq. 3) = not reported
    Points with predicted rectification magnitude above tau are discarded; no value or tuning procedure is given.
  • dr and dc (blocks allocated to Rectification and Completion Prompters) = not reported
    The split of transformer blocks between the two prompters is not specified; it affects the learned features and is presumably tuned.
  • K (number of prompt tokens in Shape-Aware Unit) = not reported
    The number of prompt tokens prepended in each block and the top-K centers assigned to them are not stated.
  • r (rank in Shape-Aware adapter) = not reported
    The rank of the low-rank adapter in Eq. 12 is not reported.
  • Staged optimization epoch schedule (20/30/50) = 20 epochs for both prompters, 30 for completion only, then downstream
    This schedule is an additional free design choice that affects training and potentially the comparison with baselines.
  • M (number of coarse centers for completion) = not reported
    The number of predicted coarse missing-part centers in Eq. 6 is not reported.
assumptions (5)
  • domain assumption Pre-trained backbone weights (Point-MAE, ReCon, Point-FEMAE) encode transferable geometric features for downstream tasks.
    The method freezes these weights; if the features are not transferable, the prompters have no useful basis. Invoked throughout Section 3.
  • domain assumption The synthetic corruption model (random viewpoint removal of 25-50% plus fixed outlier and surface noise counts) approximates real-world low-quality point clouds.
    Used to create training pairs and test sets; the ScanObjectNN experiment also applies synthetic cropping to real scans. Stated in Section 5.2 and supplementary.
  • domain assumption The displacement vector from a noisy point to its k-nearest neighbor in the clean cloud is a valid training target for rectification.
    Defines the rectification loss in Eq. 4; this geometry-based proxy assumes nearest-neighbor distances on the clean surface capture the desired denoising behavior.
  • ad hoc to paper The pre-trained MAE decoder can be repurposed, without fine-tuning, to reconstruct dense missing regions from rectified point clouds.
    Used in Eq. 6 to generate completion prompts; the decoder was trained for masked token reconstruction, not for the completion prompter's input distribution, and this transfer is assumed to work.
  • standard math Inverse-distance spatial interpolation with p=2 and top-K=6 is a sufficient feature propagation rule.
    Used in Eq. 21 (supplementary) for propagating features between centers and points; this is a standard choice, but the specific parameters are assumed.

how reviews work

0 comments
Cite this review

Pith. "Pith review of UPP: Unified Point-Level Prompting for Robust Point Cloud Analysis." pith.science (2026). https://pith.science/paper/47NK7QH5

@misc{pith2026250718997,
  author       = {Pith},
  title        = {Pith review of: UPP: Unified Point-Level Prompting for Robust Point Cloud Analysis},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/47NK7QH5}},
  note         = {Machine review of arXiv:2507.18997}
}
read the original abstract

Pre-trained point cloud analysis models have shown promising advancements in various downstream tasks, yet their effectiveness is typically suffering from low-quality point cloud (i.e., noise and incompleteness), which is a common issue in real scenarios due to casual object occlusions and unsatisfactory data collected by 3D sensors. To this end, existing methods focus on enhancing point cloud quality by developing dedicated denoising and completion models. However, due to the isolation between the point cloud enhancement and downstream tasks, these methods fail to work in various real-world domains. In addition, the conflicting objectives between denoising and completing tasks further limit the ensemble paradigm to preserve critical geometric features. To tackle the above challenges, we propose a unified point-level prompting method that reformulates point cloud denoising and completion as a prompting mechanism, enabling robust analysis in a parameter-efficient manner. We start by introducing a Rectification Prompter to adapt to noisy points through the predicted rectification vector prompts, effectively filtering noise while preserving intricate geometric features essential for accurate analysis. Sequentially, we further incorporate a Completion Prompter to generate auxiliary point prompts based on the rectified point clouds, facilitating their robustness and adaptability. Finally, a Shape-Aware Unit module is exploited to efficiently unify and capture the filtered geometric features for the downstream point cloud analysis.Extensive experiments on four datasets demonstrate the superiority and robustness of our method when handling noisy and incomplete point cloud data against existing state-of-the-art methods. Our code is released at https://github.com/zhoujiahuan1991/ICCV2025-UPP.

Figures

Figures reproduced from arXiv: 2507.18997 by the authors.

Figure 1
Figure 1. Comparison between (a) the conventional ensemble [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Our UPP pipeline processes noisy and incomplete point clouds in a unified paradigm. The input point cloud first passes [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Robustness of our method UPP and other methods [ [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Impairment of model performance by different forms of [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]
Figure 5
Figure 5. Figure 5: Visualization of Noisy ModelNet40 dataset. Our Recti [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]

Review history (2 revisions) →

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

45 extracted references · 34 canonical work pages

  1. [1]

    GAPrompt: Geometry-Aware Point Cloud Prompt for 3D Vision Model

    Zixiang Ai, Zichen Liu, Yuanhang Lei, Zhenyu Cui, Xu Zou, and Jiahuan Zhou. Gaprompt: Geometry-aware point cloud prompt for 3d vision model. arXiv preprint arXiv:2505.04119, 2025. 2, 3

    work page Pith review arXiv 2025

  2. [2]

    Vision Graph Prompting via Semantic Low-Rank Decomposition

    Zixiang Ai, Zichen Liu, and Jiahuan Zhou. Vision graph prompting via semantic low-rank decomposition. arXiv preprint arXiv:2505.04121, 2025. 3

    work page Pith review arXiv 2025

  3. [3]

    Shapenet: An information-rich 3d model repository

    Angel X Chang, Thomas Funkhouser, Leonidas Guibas, Pat Hanrahan, Qixing Huang, Zimo Li, Silvio Savarese, Manolis Savva, Shuran Song, Hao Su, et al. Shapenet: An information-rich 3d model repository. arXiv preprint arXiv:1512.03012, 2015. 6, 7

    arXiv 2015

  4. [4]

    Pointgpt: Auto-regressively generative pre- training from point clouds

    Guangyan Chen, Meiling Wang, Yi Yang, Kai Yu, Li Yuan, and Yufeng Yue. Pointgpt: Auto-regressively generative pre- training from point clouds. Advances in Neural Information Processing Systems, 36, 2024. 6

    work page 2024

  5. [5]

    Adaptformer: Adapting vision transformers for scalable visual recogni- tion

    Shoufa Chen, Chongjian Ge, Zhan Tong, Jiangliu Wang, Yibing Song, Jue Wang, and Ping Luo. Adaptformer: Adapting vision transformers for scalable visual recogni- tion. Advances in Neural Information Processing Systems , 35:16664–16678, 2022. 2

    work page 2022

  6. [6]

    Straight- pcf: Straight point cloud filtering

    Dasith de Silva Edirimuni, Xuequan Lu, Gang Li, Lei Wei, Antonio Robles-Kelly, and Hongdong Li. Straight- pcf: Straight point cloud filtering. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 20721–20730, 2024. 1, 2

    work page 2024

  7. [7]

    Runpei Dong, Zekun Qi, Linfeng Zhang, Junbo Zhang, Jian- jian Sun, Zheng Ge, Li Yi, and Kaisheng Ma. Autoencoders as cross-modal teachers: Can pretrained 2d image transform- ers help 3d representation learning? In The Eleventh Inter- national Conference on Learning Representations, 2022. 6

    work page 2022

  8. [8]

    Autoen- coders as cross-modal teachers: Can pretrained 2d image transformers help 3d representation learning? arXiv preprint arXiv:2212.08320, 2022

    Runpei Dong, Zekun Qi, Linfeng Zhang, Junbo Zhang, Jian- jian Sun, Zheng Ge, Li Yi, and Kaisheng Ma. Autoen- coders as cross-modal teachers: Can pretrained 2d image transformers help 3d representation learning? arXiv preprint arXiv:2212.08320, 2022. 2

    arXiv 2022

Show all 45 references
  1. [9]

    An image is worth 16x16 words: Transformers for image recognition at scale

    Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Syl- vain Gelly, Jakob Uszkoreit, and Neil Houlsby. An image is worth 16x16 words: Transformers for image recognition ...

    2021

  2. [10]

    T-corresnet: Template guided 3d point cloud completion with correspondence pool- ing query generation strategy

    Fan Duan, Jiahao Yu, and Li Chen. T-corresnet: Template guided 3d point cloud completion with correspondence pool- ing query generation strategy. In European Conference on Computer Vision, pages 90–106. Springer, 2025. 1, 2

    2025

  3. [11]

    Parameter-efficient transfer learning for nlp

    Neil Houlsby, Andrei Giurgiu, Stanislaw Jastrzebski, Bruna Morrone, Quentin De Laroussilhe, Andrea Gesmundo, Mona Attariyan, and Sylvain Gelly. Parameter-efficient transfer learning for nlp. In International conference on machine learning, pages 2790–2799. PMLR, 2019. 3

    2019

  4. [12]

    Lora: Low-rank adaptation of large language models

    Edward J Hu, Yelong Shen, Phillip Wallis, Zeyuan Allen- Zhu, Yuanzhi Li, Shean Wang, Lu Wang, and Weizhu Chen. Lora: Low-rank adaptation of large language models. arXiv preprint arXiv:2106.09685, 2021. 2

    2021 arXiv

  5. [13]

    Vi- sual prompt tuning

    Menglin Jia, Luming Tang, Bor-Chun Chen, Claire Cardie, Serge Belongie, Bharath Hariharan, and Ser-Nam Lim. Vi- sual prompt tuning. In European Conference on Computer Vision, pages 709–727. Springer, 2022. 3, 2

    2022

  6. [14]

    Fact: Factor-tuning for lightweight adaptation on vision transformer

    Shibo Jie and Zhi-Hong Deng. Fact: Factor-tuning for lightweight adaptation on vision transformer. In Proceed- ings of the AAAI conference on artificial intelligence , pages 1060–1068, 2023

    2023

  7. [15]

    Compacter: Efficient low-rank hypercomplex adapter layers

    Rabeeh Karimi Mahabadi, James Henderson, and Sebastian Ruder. Compacter: Efficient low-rank hypercomplex adapter layers. Advances in Neural Information Processing Systems, 34:1022–1035, 2021. 3

    2021

  8. [16]

    Oneformer3d: One transformer for unified point cloud segmentation

    Maxim Kolodiazhnyi, Anna V orontsova, Anton Konushin, and Danila Rukhovich. Oneformer3d: One transformer for unified point cloud segmentation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 20943–20953, 2024. 1

    2024

  9. [17]

    Proxyformer: Proxy alignment assisted point cloud comple- tion with missing part sensitive transformer

    Shanshan Li, Pan Gao, Xiaoyang Tan, and Mingqiang Wei. Proxyformer: Proxy alignment assisted point cloud comple- tion with missing part sensitive transformer. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 9466–9475, 2023. 1

    2023

  10. [18]

    Relation-shape convolutional neural network for point cloud analysis

    Yongcheng Liu, Bin Fan, Shiming Xiang, and Chunhong Pan. Relation-shape convolutional neural network for point cloud analysis. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages 8895– 8904, 2019. 6

    2019

  11. [19]

    Insvp: Efficient instance visual prompting from image itself

    Zichen Liu, Yuxin Peng, and Jiahuan Zhou. Insvp: Efficient instance visual prompting from image itself. In Proceedings of the 32nd ACM International Conference on Multimedia , pages 6443–6452, 2024. 3

    2024

  12. [20]

    Stop: Integrated spatial-temporal dynamic prompting for video understanding

    Zichen Liu, Kunlun Xu, Bing Su, Xu Zou, Yuxin Peng, and Jiahuan Zhou. Stop: Integrated spatial-temporal dynamic prompting for video understanding. In Proceedings of the Computer Vision and Pattern Recognition Conference, pages 13776–13786, 2025. 3

    2025

  13. [21]

    Decoupled weight de- cay regularization

    Ilya Loshchilov and Frank Hutter. Decoupled weight de- cay regularization. In International Conference on Learning Representations, 2019. 1

    2019

  14. [22]

    Sgdr: Stochastic gradient descent with warm restarts

    Ilya Loshchilov and Frank Hutter. Sgdr: Stochastic gradient descent with warm restarts. In International Conference on Learning Representations, 2022. 1

    2022

  15. [23]

    Differentiable manifold recon- struction for point cloud denoising

    Shitong Luo and Wei Hu. Differentiable manifold recon- struction for point cloud denoising. In Proceedings of the 28th ACM international conference on multimedia , pages 1330–1338, 2020. 1, 2

    2020

  16. [24]

    Score-based point cloud denoising

    Shitong Luo and Wei Hu. Score-based point cloud denoising. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 4583–4592, 2021. 1, 2, 6

    2021

  17. [25]

    Rethinking network design and local geometry in point 9 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 9 cloud: A simple residual mlp framework. arXiv preprint arXiv:2202.07123, 2022. 6

    2022 arXiv

  18. [26]

    Masked autoencoders for point cloud self-supervised learning

    Yatian Pang, Wenxiao Wang, Francis EH Tay, Wei Liu, Yonghong Tian, and Li Yuan. Masked autoencoders for point cloud self-supervised learning. In European conference on computer vision, pages 604–621. Springer, 2022. 1, 2, 4, 5, 6, 7, 8

    2022

  19. [27]

    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,

  20. [28]

    Contrast with reconstruct: Contrastive 3d representation learning guided by generative pretraining

    Zekun Qi, Runpei Dong, Guofan Fan, Zheng Ge, Xiangyu Zhang, Kaisheng Ma, and Li Yi. Contrast with reconstruct: Contrastive 3d representation learning guided by generative pretraining. In International Conference on Machine Learn- ing, pages 28223–28243. PMLR, 2023. 4, 5, 6, 7, 1

    2023

  21. [29]

    Shapellm: Universal 3d object understanding for embodied interaction

    Zekun Qi, Runpei Dong, Shaochen Zhang, Haoran Geng, Chunrui Han, Zheng Ge, Li Yi, and Kaisheng Ma. Shapellm: Universal 3d object understanding for embodied interaction. arXiv preprint arXiv:2402.17766, 2024. 2

    2024 arXiv

  22. [30]

    Pointcleannet: Learning to denoise and remove outliers from dense point clouds

    Marie-Julie Rakotosaona, Vittorio La Barbera, Paul Guer- rero, Niloy J Mitra, and Maks Ovsjanikov. Pointcleannet: Learning to denoise and remove outliers from dense point clouds. In Computer graphics forum, pages 185–203. Wiley Online Library, 2020. 2

    2020

  23. [31]

    Contrastive boundary learning for point cloud segmentation

    Liyao Tang, Yibing Zhan, Zhe Chen, Baosheng Yu, and Dacheng Tao. Contrastive boundary learning for point cloud segmentation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 8489–8499, 2022. 1

    2022

  24. [32]

    Point- peft: Parameter-efficient fine-tuning for 3d pre-trained mod- els

    Yiwen Tang, Ray Zhang, Zoey Guo, Xianzheng Ma, Bin Zhao, Zhigang Wang, Dong Wang, and Xuelong Li. Point- peft: Parameter-efficient fine-tuning for 3d pre-trained mod- els. In Proceedings of the AAAI Conference on Artificial In- telligence, pages 5171–5179, 2024. 2, 3, 6, 7, 1

    2024

  25. [33]

    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

  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. 6, 1, 2

    1912

  27. [35]

    Compo- nential prompt-knowledge alignment for domain incremen- tal learning

    Kunlun Xu, Xu Zou, Gang Hua, and Jiahuan Zhou. Compo- nential prompt-knowledge alignment for domain incremen- tal learning. arXiv preprint arXiv:2505.04575, 2025. 3

    2025 arXiv

  28. [36]

    Pointasnl: Robust point clouds processing using nonlocal neural networks with adaptive sampling

    Xu Yan, Chaoda Zheng, Zhen Li, Sheng Wang, and Shuguang Cui. Pointasnl: Robust point clouds processing using nonlocal neural networks with adaptive sampling. In Proceedings of the IEEE/CVF conference on computer vi- sion and pattern recognition, pages 5589–5598, 2020. 6

    2020

  29. [37]

    A scalable active framework for region annotation in 3d shape collections.ACM Transactions on Graphics (ToG), 35(6):1–12, 2016

    Li Yi, Vladimir G Kim, Duygu Ceylan, I-Chao Shen, Mengyan Yan, Hao Su, Cewu Lu, Qixing Huang, Alla Shef- fer, and Leonidas Guibas. A scalable active framework for region annotation in 3d shape collections.ACM Transactions on Graphics (ToG), 35(6):1–12, 2016. 1

    2016

  30. [38]

    Pointr: Diverse point cloud comple- tion with geometry-aware transformers

    Xumin Yu, Yongming Rao, Ziyi Wang, Zuyan Liu, Jiwen Lu, and Jie Zhou. Pointr: Diverse point cloud comple- tion with geometry-aware transformers. In Proceedings of the IEEE/CVF international conference on computer vision, pages 12498–12507, 2021. 1, 2, 6

    2021

  31. [39]

    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, 6

    2022

  32. [40]

    Instance-aware dynamic prompt tuning for pre-trained point cloud models

    Yaohua Zha, Jinpeng Wang, Tao Dai, Bin Chen, Zhi Wang, and Shu-Tao Xia. Instance-aware dynamic prompt tuning for pre-trained point cloud models. In Proceedings of the IEEE/CVF International Conference on Computer Vision , pages 14161–14170, 2023. 2, 3, 7, 1

    2023

  33. [41]

    Towards compact 3d representations via point feature enhancement masked au- toencoders

    Yaohua Zha, Huizhen Ji, Jinmin Li, Rongsheng Li, Tao Dai, Bin Chen, Zhi Wang, and Shu-Tao Xia. Towards compact 3d representations via point feature enhancement masked au- toencoders. In Proceedings of the AAAI Conference on Ar- tificial Intelligence, pages 6962–6970, 2024. 1, ...

    2024

  34. [42]

    Point-m2ae: multi-scale masked autoencoders for hierarchical point cloud pre-training

    Renrui Zhang, Ziyu Guo, Peng Gao, Rongyao Fang, Bin Zhao, Dong Wang, Yu Qiao, and Hongsheng Li. Point-m2ae: multi-scale masked autoencoders for hierarchical point cloud pre-training. Advances in neural information processing sys- tems, 35:27061–27074, 2022. 2

    2022

  35. [43]

    Pcp- mae: Learning to predict centers for point masked autoen- coders

    Xiangdong Zhang, Shaofeng Zhang, and Junchi Yan. Pcp- mae: Learning to predict centers for point masked autoen- coders. Advances in Neural Information Processing Systems, 37:80303–80327, 2025. 6

    2025

  36. [44]

    Dynamic adapter meets prompt tuning: Parameter-efficient transfer learning for point cloud analysis

    Xin Zhou, Dingkang Liang, Wei Xu, Xingkui Zhu, Yihan Xu, Zhikang Zou, and Xiang Bai. Dynamic adapter meets prompt tuning: Parameter-efficient transfer learning for point cloud analysis. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages...

    2024

  37. [45]

    Point- clip v2: Prompting clip and gpt for powerful 3d open-world learning

    Xiangyang Zhu, Renrui Zhang, Bowei He, Ziyu Guo, Ziyao Zeng, Zipeng Qin, Shanghang Zhang, and Peng Gao. Point- clip v2: Prompting clip and gpt for powerful 3d open-world learning. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 2639–2650, 2023...

    2023

Pith tools

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