Pith. sign in

REVIEW 4 major objections 4 minor 1 cited by

IPDN: Image-enhanced Prompt Decoding Network for 3D Referring Expression Segmentation

T0 review · 4 major / 4 minor · reviewed 2026-08-10 · deepseek-v4-flash

Pith's one-line read IPDN beats prior 3D-RES models by 1.9 and 4.2 mIoU points by injecting CLIP image features and task-driven prompts into the decoder.

desk verdict A sensible incremental combination of CLIP multi-view fusion and thresholded prompt decoding for 3D referring segmentation, with plausible single-run gains that need repeated-seed confirmation. read the letter →

arxiv 2501.04995 v1 pith:ICKEK2ET submitted 2025-01-09 cs.CV cs.AI

classification cs.CVcs.AI
keywords 3Dreferringexpressionsegmentationpointcloudmulti-viewsemanticembeddingprompt-awaredecodertask-drivenpromptsCLIPScanReferMulti3DRefer
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper claims that the two known weaknesses of 3D referring expression segmentation—features computed from point clouds alone are noisy and under-aligned with language, and decoders treat every object query as equally relevant—can be fixed by two targeted modules. The proposed IPDN injects multi-view CLIP image features into the point cloud via geometric projection to repair lost or distorted visual information, and adds a prompt-aware decoder that derives task-driven guidance from cross-attention between text and queries. The authors report that on standard benchmarks this raises mean Intersection-over-Union by 1.9 points for 3D-RES and by 4.2 points for 3D-GRES, with the largest gains on rare object classes and on scenes cluttered with distractors. If those numbers hold, a practical takeaway is that using cheap 2D pretrained features plus a text-relevance signal is a reliable way to make 3D segmentation interpret language.

What carries the argument

The two load-bearing mechanisms are the Multi-view Semantic Embedding (MSE) module and the Prompt-Aware Decoder (PAD). MSE takes CLIP patch features upsampled to pixel level, projects each pixel into 3D via the camera intrinsics, extrinsics, and depth (Eq. 5), and uses spherical querying to attach the projected 2D features to nearby points, then adds a spatial-aware masked self-attention over k-nearest neighbor queries so the fused features keep explicit 3D relations. PAD runs cross-attention with text as query and sparse queries as keys and values, sums the attention scores to get per-query relevance, filters out low-confidence queries by thresholding with the previous layer's predicted probability, softmaxes the rest, and concatenates the resulting task-driven prompt to the queries before each fusion layer. The prediction head then selects the highest-probability query (3D-RES) or merges all queries with probability above 0.5 (3D-GRES).

What would settle it

Perturb the camera intrinsics and extrinsics or add realistic depth noise to the ScanRefer validation views and re-measure IPDN's mIoU; if the gain over a point-cloud-only baseline vanishes or turns negative under calibration errors within sensor tolerance, the central claim of the MSE module is not robust.

Watch

Extended reading notes

Core claim

The paper's central claim is that both feature ambiguity and intent ambiguity in 3D referring expression segmentation are reducible to two architectural interventions. By projecting CLIP's pixel-level image features into the point cloud with known camera parameters and averaging over overlapping views, the MSE module produces a fused visual representation that aligns better with text and preserves spatial structure; by computing cross-attention scores between text and sparse object queries, thresholding them with the model's own confidence, and attaching the weighted queries as prompts, the PAD module steers the decoder toward text-relevant queries. The paper's controlled ablation attributes roughly equal gains to the two modules (1.1 and 1.0 mIoU points) and reports state-of-the-art results on ScanRefer and Multi3DRefer.

Load-bearing premise

The MSE module assumes every training and test view comes with accurate camera intrinsics, extrinsics, and per-pixel depth, since any error in the projection of Eq. (5) degrades the image features injected into the point cloud.

Editorial extensions

If this is right

  • On the ScanRefer benchmark, IPDN reports 60.6 Acc@0.25, 54.9 Acc@0.5, and 50.2 mIoU, ahead of the previous best MDIN by 2.6, 1.8, and 1.9 points respectively.
  • On Multi3DRefer, IPDN reports 51.7 mIoU overall, 4.2 points above MDIN, with the biggest single-target gain at Acc@0.5 (8 points without distractors).
  • The model's mIoU drops only 6.3 points from high-frequency to low-frequency object classes, versus 17.5 for 3D-STMN and 14.9 for MDIN, suggesting the 2D-pretrained features help rare categories.
  • Ablations show the two modules are approximately additive: MSE contributes 1.1 mIoU and PAD contributes 1.0 mIoU over the baseline without either.

Reading between the lines

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

  • A testable extension is to run IPDN on RGB-D scans whose camera poses or depth maps are synthetically perturbed; if the mIoU gain over point-cloud-only baselines shrinks under realistic calibration noise, the geometric projection in MSE is the fragile link.
  • The same recipe of CLIP feature injection plus confidence-gated prompt decoding could transfer to 2D referring segmentation or open-vocabulary 3D instance segmentation, where feature ambiguity and query ranking also matter.
  • Because the paper only evaluates on ScanNet-derived benchmarks, the performance on noisy outdoor or multi-room scans, or scenes with missing views, remains unknown; the claimed gains may not extrapolate until the sensitivity to view coverage is measured.
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 / 4 minor

Summary. The paper proposes IPDN, a one-stage query-based network for 3D Referring Expression Segmentation (3D-RES) and Generalized 3D Referring Expression Segmentation (3D-GRES). It introduces two modules: MSE (Multi-view Semantic Embedding), which projects CLIP-extracted 2D image features into the point cloud via camera parameters and integrates them with 3D superpoint features under a spatial-aware attention mechanism, and PAD (Prompt-aware Decoder), which generates task-driven prompts from cross-attention scores between text and queries, filtering irrelevant queries using predicted probabilities. Experiments on ScanRefer and Multi3DRefer report improvements over the previous state of the art MDIN by 1.9 mIoU on 3D-RES and 4.2 mIoU on 3D-GRES, with ablations showing each module contributes roughly 1 point of mIoU.

Significance. If the reported gains are statistically robust, IPDN is a meaningful advance: it combines large-scale 2D vision-language pretraining with a lightweight task-driven prompt mechanism, and the long-tail analysis in Table 3 shows a substantially smaller performance drop on rare classes than prior 3D-only methods, which is a credible benefit of injecting CLIP features. The paper is clearly motivated, the component ablations are internally consistent, and the authors release code, which aids reproducibility. However, the current evidence is weakened by single-run evaluations without error bars or significance tests, and several implementation details needed for faithful reproduction or for assessing generality are missing. The significance is therefore contingent on addressing these empirical and specification gaps.

major comments (4)
  1. [Sec. 4.3, Tables 2 and 4] The central claim of state-of-the-art performance rests on single-run comparisons. IPDN reports 50.2 overall mIoU versus MDIN's 48.3 on ScanRefer, but the ablation without MSE and PAD already reaches 48.1, so the 1.9-point margin over the previous SOTA is essentially the sum of two roughly 1-point module gains. On a test set of this size, 1–2 mIoU differences are commonly within run-to-run variance, and no error bars, number of seeds, or significance tests are reported. Please provide mean and standard deviation over at least three runs for the main comparisons and the key ablations, or otherwise justify that the reported differences are not noise.
  2. [Sec. 3.3, Eq. (8)] The filtering operation in Eq. (8) uses P rob^{l−1}, but P rob^0 is never defined for the first decoder layer. Since PAD is one of the two main contributions, the paper must specify how the first-layer probabilities are initialized (for example, uniform values, a prediction from Q0, or skipping filtering at l=1), and confirm that the reported ablations use that exact choice.
  3. [Sec. 3.2] The spherical querying step in the MSE module is underspecified: the sphere radius is not reported, the aggregation over multiple views is described only as an average, and the number of input views N_I and how they are selected are not given in Section 4.1. These choices directly control how much 2D information enters the 3D features and could materially affect the results; they need to be stated explicitly, ideally with a sensitivity analysis.
  4. [Sec. 3.2, Eq. (5), Sec. 4.2] MSE assumes accurate camera intrinsics, extrinsics, and per-pixel depth for every input view, and the evaluation is limited to ScanNet-based benchmarks that provide this metadata. The paper does not test sensitivity to calibration error, missing views, or imperfect depth, and it does not discuss behavior on data without such metadata. Please add a robustness experiment or explicitly bound the generality claim so that the reported gains are not over-interpreted as holding in settings where the required camera information is unavailable.
minor comments (4)
  1. [Section 4 heading and Table 1] The section heading reads "Expriments" and should be "Experiments"; Table 1's header duplicates "Acc@0.25 Acc@0.5" without a separating line, which makes the ZT/ST/MT columns hard to parse.
  2. [Eq. (5)] The projection formula uses R(K^{-1}[u,v,1]^T D) + T; please clarify the coordinate convention (e.g., whether R is the rotation from camera to world or its inverse) so that the projection can be reproduced exactly.
  3. [Sec. 4.1] The learning-rate schedule is called "PolyRL" and is likely a typo for "PolyLR"; please correct it and state the optimizer and weight decay.
  4. [Sec. 4.4 and throughout] There are several small language issues: "r is not necessarily better when it is larger" should be "a larger r is not always better"; "P robvalues" needs a space; and the abstract's "state-ofthe-art" should be "state-of-the-art."

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: IPDN's gains are empirical comparisons against published baselines, and the self-referential prompt mechanism is supervised architectural feedback, not definitional.

full rationale

The central claim is an empirical SOTA improvement on ScanRefer and Multi3DRefer. The MSE module (Eq. 5, spherical querying) injects CLIP features into the point cloud via known camera geometry; the PAD module (Eqs. 7-10) generates prompts from the model's own cross-attention scores and probability estimates. This is a recurrent, end-to-end-trained feedback loop, not a definitional equivalence: the prompts do not encode the reported mIoU or ground-truth masks, and the loss (Eqs. 15-17) provides independent supervision. The comparisons to MDIN and 3D-STMN use published, externally benchmarked results from the same group; citing and building on one's own decoder is standard incremental work, not a load-bearing self-citation chain. Concerns about missing P rob^0 initialization, unreported sphere radius, single-run results, and tuning k/r on the evaluation benchmark are reproducibility and statistical issues, not circularity by construction. No equation in the paper reduces a predicted quantity to a fitted input or to the target metric.

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

No new physical entity is introduced. The method relies on external pretrained models (CLIP, RoBERTa), prior 3D modules (MDIN, EDA), and accurate camera geometry. The main fitted quantities are hyperparameters r and k, selected by ablations, plus an unstated spherical radius, which makes the contribution partially dependent on tuning choices rather than on independent evidence.

free parameters (4)
  • probability threshold r = 0.75
    Threshold in the Prompt-Aware Decoder, chosen via ablation in Table 6 to maximize mIoU.
  • nearest-neighbor count k = 8
    Neighbor count in Spatial-aware Attention, chosen via ablation in Table 5 to maximize mIoU.
  • spherical query radius = unknown
    Radius used to assign 2D pixel features to nearby 3D points in Sec. 3.2; no value is reported or ablated.
  • loss weights lambda_b, lambda_p, lambda_c = 1.0, 0.1, 0.1
    Hand-set weights in Eq. (17); they are not ablated or justified beyond being conventional choices.
assumptions (4)
  • domain assumption Camera intrinsics, extrinsics, and depth values are accurate for every input view.
    Eq. (5) projects 2D pixel features into the 3D point cloud using these parameters; any calibration error misplaces the injected features.
  • domain assumption CLIP visual features align with textual concepts and transfer to indoor object classes.
    The MSE module relies on CLIP pretraining to enrich point cloud features, stated in Sec. 3.2; if CLIP cannot represent indoor objects well, the fusion adds noise.
  • domain assumption Cross-attention scores between text and queries reflect true query relevance.
    The PAD module uses these scores to build task-driven prompts in Eqs. (7)-(9); the whole mechanism assumes the scores are a meaningful relevance signal.
  • domain assumption The MDIN sampling module, feature fusion, and EDA contrastive loss are reproducible and effective.
    The decoder adopts MDIN's sampling and feature fusion (Eq. (11)) and EDA's contrastive loss in Sec. 3.4, with details outsourced to citations from the same research group.

how reviews work

0 comments
Cite this review

Pith. "Pith review of IPDN: Image-enhanced Prompt Decoding Network for 3D Referring Expression Segmentation." pith.science (2026). https://pith.science/paper/ICKEK2ET

@misc{pith2026250104995,
  author       = {Pith},
  title        = {Pith review of: IPDN: Image-enhanced Prompt Decoding Network for 3D Referring Expression Segmentation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ICKEK2ET}},
  note         = {Machine review of arXiv:2501.04995}
}
read the original abstract

3D Referring Expression Segmentation (3D-RES) aims to segment point cloud scenes based on a given expression. However, existing 3D-RES approaches face two major challenges: feature ambiguity and intent ambiguity. Feature ambiguity arises from information loss or distortion during point cloud acquisition due to limitations such as lighting and viewpoint. Intent ambiguity refers to the model's equal treatment of all queries during the decoding process, lacking top-down task-specific guidance. In this paper, we introduce an Image enhanced Prompt Decoding Network (IPDN), which leverages multi-view images and task-driven information to enhance the model's reasoning capabilities. To address feature ambiguity, we propose the Multi-view Semantic Embedding (MSE) module, which injects multi-view 2D image information into the 3D scene and compensates for potential spatial information loss. To tackle intent ambiguity, we designed a Prompt-Aware Decoder (PAD) that guides the decoding process by deriving task-driven signals from the interaction between the expression and visual features. Comprehensive experiments demonstrate that IPDN outperforms the state-ofthe-art by 1.9 and 4.2 points in mIoU metrics on the 3D-RES and 3D-GRES tasks, respectively.

Figures

Figures reproduced from arXiv: 2501.04995 by the authors.

Figure 1
Figure 1. The pipeline of (a) the previous traditional query [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. The overview of our framework. denotes the number of tokens, and Ct indicates the Ct￾dimensionality of each embedding. In order to have a unified feature dimension d in the decoder, we transform E into tex￾tual features T ∈ R Nt×d via a linear projection: T = EWt, (1) where Wt ∈ R Ct×d are learnable parameters. Visual Feature Given a point cloud scene P ∈ R Np×(3+f) , where Np denotes the number of points. Each poin… view at source ↗
Figure 3
Figure 3. Qualitative comparison between MDIN and ours. [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗

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. 3DResT: A Strong Baseline for Semi-Supervised 3D Referring Expression Segmentation

    cs.CV 2025-04 conditional novelty 6.0 of 10

    3DResT applies teacher-student semi-supervised learning to 3D referring expression segmentation, promoting high-agreement pseudo-labels into the labeled set and dynamically weighting low-agreement ones, and reports a ...

Reference graph

Works this paper leans on

76 extracted references · 72 canonical work pages · cited by 1 Pith paper

  1. [1]

    Z.; Chang, A

    Chen, D. Z.; Chang, A. X.; and Nie ner, M. 2020. Scanrefer: 3d object localization in rgb-d scans using natural language. In ECCV

  2. [2]

    Chen, S.; Guhur, P.-L.; Tapaswi, M.; Schmid, C.; and Laptev, I. 2022. Language conditioned spatial relation reasoning for 3d object grounding. NeurIPS

  3. [3]

    X.; Zheng, H.; Han, Y.; Qiu, X.; and Huang, G

    Chng, Y. X.; Zheng, H.; Han, Y.; Qiu, X.; and Huang, G. 2024. Mask grounding for referring image segmentation. In CVPR

  4. [4]

    X.; Savva, M.; Halber, M.; Funkhouser, T.; and Nie ner, M

    Dai, A.; Chang, A. X.; Savva, M.; Halber, M.; Funkhouser, T.; and Nie ner, M. 2017. Scannet: Richly-annotated 3d reconstructions of indoor scenes. In CVPR

  5. [5]

    Deng, J.; Shi, S.; Li, P.; Zhou, W.; Zhang, Y.; and Li, H. 2021. Voxel r-cnn: Towards high performance voxel-based 3d object detection. In AAAI

  6. [6]

    Ding, H.; Liu, C.; Wang, S.; and Jiang, X. 2021. Vision-language transformer and query generation for referring segmentation. In ICCV

  7. [7]

    Dong, R.; Qi, Z.; Zhang, L.; Zhang, J.; Sun, J.; Ge, Z.; Yi, L.; and Ma, K. 2022. Autoencoders as cross-modal teachers: Can pretrained 2d image transformers help 3d representation learning? arXiv

  8. [8]

    Fei, H.; Wu, S.; Ji, W.; Zhang, H.; Zhang, M.; Lee, M.-L.; and Hsu, W. 2024 a . Video-of-thought: Step-by-step video reasoning from perception to cognition. In ICML

Show all 76 references
  1. [9]

    Fei, H.; Wu, S.; Zhang, H.; Chua, T.-S.; and Shuicheng, Y. 2024 b . Vitron: A unified pixel-level vision llm for understanding, generating, segmenting, editing. In NeurIPS

  2. [10]

    Fei, H.; Wu, S.; Zhang, M.; Zhang, M.; Chua, T.-S.; and Yan, S. 2024 c . Enhancing video-language representations with structural spatio-temporal alignment. TPAMI

  3. [11]

    Feng, M.; Li, Z.; Li, Q.; Zhang, L.; Zhang, X.; Zhu, G.; Zhang, H.; Wang, Y.; and Mian, A. 2021. Free-form description guided 3d visual graph network for object grounding in point cloud. In ICCV

  4. [12]

    Graham, B.; Engelcke, M.; and Van Der Maaten, L. 2018. 3d semantic segmentation with submanifold sparse convolutional networks. In CVPR

  5. [13]

    He, D.; Zhao, Y.; Luo, J.; Hui, T.; Huang, S.; Zhang, A.; and Liu, S. 2021. Transrefer3d: Entity-and-relation aware transformer for fine-grained 3d visual grounding. In ACM MM

  6. [14]

    He, S.; and Ding, H. 2024. RefMask3D: Language-Guided Transformer for 3D Referring Segmentation. arXiv

  7. [15]

    He, S.; Ding, H.; Jiang, X.; and Wen, B. 2024. SegPoint: Segment Any Point Cloud via Large Language Model. arXiv

  8. [16]

    Huang, K.-C.; Li, X.; Qi, L.; Yan, S.; and Yang, M.-H. 2024. Reason3D: Searching and Reasoning 3D Segmentation via Large Language Model. arXiv

  9. [17]

    Huang, P.-H.; Lee, H.-H.; Chen, H.-T.; and Liu, T.-L. 2021. Text-guided graph neural networks for referring 3d instance segmentation. In AAAI

  10. [18]

    Jia, M.; Tang, L.; Chen, B.-C.; Cardie, C.; Belongie, S.; Hariharan, B.; and Lim, S.-N. 2022. Visual prompt tuning. In ECCV

  11. [19]

    Kolodiazhnyi, M.; Vorontsova, A.; Konushin, A.; and Rukhovich, D. 2024. Oneformer3d: One transformer for unified point cloud segmentation. In CVPR

  12. [20]

    Lai, X.; Liu, J.; Jiang, L.; Wang, L.; Zhao, H.; Liu, S.; Qi, X.; and Jia, J. 2022. Stratified transformer for 3d point cloud segmentation. In CVPR

  13. [21]

    Lai, X.; Tian, Z.; Chen, Y.; Li, Y.; Yuan, Y.; Liu, S.; and Jia, J. 2024. Lisa: Reasoning segmentation via large language model. In CVPR

  14. [22]

    Lai, X.; Yuan, Y.; Chu, R.; Chen, Y.; Hu, H.; and Jia, J. 2023. Mask-attention-free transformer for 3d instance segmentation. In ICCV

  15. [23]

    Landrieu, L.; and Simonovsky, M. 2018. Large-scale point cloud semantic segmentation with superpoint graphs. In CVPR

  16. [24]

    Lester, B.; Al-Rfou, R.; and Constant, N. 2021. The power of scale for parameter-efficient prompt tuning. arXiv

  17. [25]

    L.; and Liang, P

    Li, X. L.; and Liang, P. 2021. Prefix-tuning: Optimizing continuous prompts for generation. arXiv

  18. [26]

    Lin, H.; Luo, Y.; Zheng, X.; Li, L.; Chao, F.; Jin, T.; Luo, D.; Wang, C.; Wang, Y.; and Cao, L. 2023. A unified framework for 3d point cloud visual grounding. arXiv

  19. [27]

    Liu, C.; Ding, H.; and Jiang, X. 2023. Gres: Generalized referring expression segmentation. In CVPR

  20. [28]

    L.; Du, Z.; Yang, Z.; and Tang, J

    Liu, X.; Ji, K.; Fu, Y.; Tam, W. L.; Du, Z.; Yang, Z.; and Tang, J. 2021. P-tuning v2: Prompt tuning can be comparable to fine-tuning universally across scales and tasks. arXiv

  21. [29]

    Liu, Y.; Ott, M.; Goyal, N.; Du, J.; Joshi, M.; Chen, D.; Levy, O.; Lewis, M.; Zettlemoyer, L.; and Stoyanov, V. 2019. Roberta: A robustly optimized bert pretraining approach. arXiv

  22. [30]

    Lu, J.; Deng, J.; Wang, C.; He, J.; and Zhang, T. 2023 a . Query refinement transformer for 3d instance segmentation. In ICCV

  23. [31]

    P.; Boularias, A.; and Bekris, K

    Lu, S.; Chang, H.; Jing, E. P.; Boularias, A.; and Bekris, K. 2023 b . Ovir-3d: Open-vocabulary 3d instance retrieval without training on 3d data. In CoRL

  24. [32]

    Luo, J.; Fu, J.; Kong, X.; Gao, C.; Ren, H.; Shen, H.; Xia, H.; and Liu, S. 2022. 3d-sps: Single-stage 3d visual grounding via referred point progressive selection. In CVPR

  25. [33]

    Milletari, F.; Navab, N.; and Ahmadi, S.-A. 2016. V-net: Fully convolutional neural networks for volumetric medical image segmentation. In 3DV

  26. [34]

    Moenning, C.; and Dodgson, N. A. 2003. Fast marching farthest point sampling. Technical report, University of Cambridge, Computer Laboratory

  27. [35]

    Peng, S.; Genova, K.; Jiang, C.; Tagliasacchi, A.; Pollefeys, M.; Funkhouser, T.; et al. 2023. Openscene: 3d scene understanding with open vocabularies. In CVPR

  28. [36]

    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 CVPR

  29. [37]

    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. NeurIPS

  30. [38]

    Qian, Z.; Ma, Y.; Ji, J.; and Sun, X. 2024 a . X-RefSeg3D: Enhancing Referring 3D Instance Segmentation via Structured Cross-Modal Graph Neural Networks. In AAAI

  31. [39]

    Qian, Z.; Ma, Y.; Lin, Z.; Ji, J.; Zheng, X.; Sun, X.; and Ji, R. 2024 b . Multi-branch Collaborative Learning Network for 3D Visual Grounding. arXiv

  32. [40]

    W.; Hallacy, C.; Ramesh, A.; Goh, G.; Agarwal, S.; Sastry, G.; Askell, A.; Mishkin, P.; Clark, J.; et al

    Radford, A.; Kim, J. W.; Hallacy, C.; Ramesh, A.; Goh, G.; Agarwal, S.; Sastry, G.; Askell, A.; Mishkin, P.; Clark, J.; et al. 2021. Learning transferable visual models from natural language supervision. In ICML

  33. [41]

    Roh, J.; Desingh, K.; Farhadi, A.; and Fox, D. 2022. Languagerefer: Spatial-language model for 3d visual grounding. In CoRL

  34. [42]

    Rozenberszki, D.; Litany, O.; and Dai, A. 2022. Language-grounded indoor 3d semantic segmentation in the wild. In ECCV

  35. [43]

    Schult, J.; Engelmann, F.; Hermans, A.; Litany, O.; Tang, S.; and Leibe, B. 2023. Mask3d: Mask transformer for 3d semantic instance segmentation. In ICRA

  36. [44]

    A.; VS, V.; and Patel, V

    Shah, N. A.; VS, V.; and Patel, V. M. 2024. LQMFormer: Language-aware Query Mask Transformer for Referring Image Segmentation. In CVPR

  37. [45]

    Shi, S.; Guo, C.; Jiang, L.; Wang, Z.; Shi, J.; Wang, X.; and Li, H. 2020. Pv-rcnn: Point-voxel feature set abstraction for 3d object detection. In CVPR

  38. [46]

    Sun, J.; Qing, C.; Tan, J.; and Xu, X. 2023. Superpoint transformer for 3d scene instance segmentation. In AAAI

  39. [47]

    W.; Pollefeys, M.; Tombari, F.; and Engelmann, F

    Takmaz, A.; Fedele, E.; Sumner, R. W.; Pollefeys, M.; Tombari, F.; and Engelmann, F. 2023. Openmask3d: Open-vocabulary 3d instance segmentation. arXiv

  40. [48]

    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. NeurIPS

  41. [49]

    Wang, Z.; Huang, H.; Zhao, Y.; Li, L.; Cheng, X.; Zhu, Y.; Yin, A.; and Zhao, Z. 2023. 3drp-net: 3d relative position-aware network for 3d visual grounding. arXiv

  42. [50]

    Wang, Z.; Li, Y.; Liu, T.; Zhao, H.; and Wang, S. 2024. OV-Uni3DETR: Towards Unified Open-Vocabulary 3D Object Detection via Cycle-Modality Propagation. arXiv

  43. [51]

    Wang, Z.; Lu, Y.; Li, Q.; Tao, X.; Guo, Y.; Gong, M.; and Liu, T. 2022 a . Cris: Clip-driven referring image segmentation. In CVPR

  44. [52]

    Wang, Z.; Zhang, Z.; Lee, C.-Y.; Zhang, H.; Sun, R.; Ren, X.; Su, G.; Perot, V.; Dy, J.; and Pfister, T. 2022 b . Learning to prompt for continual learning. In CVPR

  45. [53]

    Wu, C.; Liu, Y.; Ji, J.; Ma, Y.; Wang, H.; Luo, G.; Ding, H.; Sun, X.; and Ji, R. 2024 a . 3D-GRES: Generalized 3D Referring Expression Segmentation. arXiv

  46. [54]

    Wu, C.; Ma, Y.; Chen, Q.; Wang, H.; Luo, G.; Ji, J.; and Sun, X. 2024 b . 3d-stmn: Dependency-driven superpoint-text matching network for end-to-end 3d referring expression segmentation. In AAAI

  47. [55]

    Wu, T.-Y.; Huang, S.-Y.; and Wang, Y.-C. F. 2024. DOrA: 3D Visual Grounding with Order-Aware Referring. arXiv

  48. [56]

    Wu, Y.; Cheng, X.; Zhang, R.; Cheng, Z.; and Zhang, J. 2023. Eda: Explicit text-decoupling and dense alignment for 3d visual grounding. In CVPR

  49. [57]

    Xu, W.; Shi, C.; Tu, S.; Zhou, X.; Liang, D.; and Bai, X. 2024. A Unified Framework for 3D Scene Understanding. arXiv

  50. [58]

    Yan, M.; Zhang, J.; Zhu, Y.; and Wang, H. 2024. Maskclustering: View consensus based mask graph clustering for open-vocabulary 3d instance segmentation. In CVPR

  51. [59]

    Yang, D.; Ji, J.; Ma, Y.; Guo, T.; Wang, H.; Sun, X.; and Ji, R. 2024 a . SAM as the Guide: Mastering Pseudo-Label Refinement in Semi-Supervised Referring Expression Segmentation. arXiv

  52. [60]

    Yang, L.; Zhang, Z.; Qi, Z.; Xu, Y.; Liu, W.; Shan, Y.; Li, B.; Yang, W.; Li, P.; Wang, Y.; et al. 2024 b . Exploiting contextual objects and relations for 3d visual grounding. NeurIPS

  53. [61]

    Yang, Y.-Q.; Guo, Y.-X.; Xiong, J.-Y.; Liu, Y.; Pan, H.; Wang, P.-S.; Tong, X.; and Guo, B. 2023. Swin3d: A pretrained transformer backbone for 3d indoor scene understanding. arXiv

  54. [62]

    Yang, Z.; Wang, J.; Tang, Y.; Chen, K.; Zhao, H.; and Torr, P. H. 2022. Lavt: Language-aware vision transformer for referring image segmentation. In CVPR

  55. [63]

    Yang, Z.; Zhang, S.; Wang, L.; and Luo, J. 2021. Sat: 2d semantics assisted training for 3d visual grounding. In ICCV

  56. [64]

    Yu, Q.; Du, H.; Liu, C.; and Yu, X. 2024. When 3D Bounding-Box Meets SAM: Point Cloud Instance Segmentation with Weak-and-Noisy Supervision. In WACV

  57. [65]

    Yuan, Z.; Yan, X.; Liao, Y.; Zhang, R.; Wang, S.; Li, Z.; and Cui, S. 2021. Instancerefer: Cooperative holistic understanding for visual grounding on point clouds through instance multi-level contextual referring. In ICCV

  58. [66]

    Zhang, J.; Dong, R.; and Ma, K. 2023. Clip-fo3d: Learning free open-world 3d scene representations from 2d dense clip. In ICCV

  59. [67]

    Zhang, J.; Fan, G.; Wang, G.; Su, Z.; Ma, K.; and Yi, L. 2023. Language-assisted 3D feature learning for semantic scene understanding. In AAAI

  60. [68]

    Zhang, Y.; Gong, Z.; and Chang, A. X. 2023. Multi3drefer: Grounding text description to multiple 3d objects. In ICCV

  61. [69]

    Zhang, Y.; Zhou, K.; and Liu, Z. 2022. Neural prompt search. arXiv

  62. [70]

    H.; and Koltun, V

    Zhao, H.; Jiang, L.; Jia, J.; Torr, P. H.; and Koltun, V. 2021 a . Point transformer. In ICCV

  63. [71]

    Zhao, L.; Cai, D.; Sheng, L.; and Xu, D. 2021 b . 3DVG-Transformer: Relation modeling for visual grounding on point clouds. In ICCV

  64. [72]

    C.; and Liu, Z

    Zhou, K.; Yang, J.; Loy, C. C.; and Liu, Z. 2022 a . Conditional prompt learning for vision-language models. In CVPR

  65. [73]

    C.; and Liu, Z

    Zhou, K.; Yang, J.; Loy, C. C.; and Liu, Z. 2022 b . Learning to prompt for vision-language models. IJCV

  66. [74]

    Zhu, B.; Niu, Y.; Han, Y.; Wu, Y.; and Zhang, H. 2023. Prompt-aligned gradient for prompt tuning. In ICCV

  67. [75]

    , " * 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...

  68. [76]

    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 gl...

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.