Pith. sign in

REVIEW 4 major objections 4 minor 2 cited by

Blended Point Cloud Diffusion for Localized Text-guided Shape Editing

T0 review · 4 major / 4 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read BlendedPC reframes localized text-guided point cloud editing as semantic inpainting, adding an inversion-free coordinate blending step that preserves the original shape's identity, and reports it outperforming prior editing methods on…

desk verdict Solid empirical systems paper: localized text-guided point-cloud editing that actually works, with a clever inversion-free blending trick; main risk is an unquantified index-consistency assumption, but the paper earns referee time. read the letter →

arxiv 2507.15399 v1 pith:ABITGGX2 submitted 2025-07-21 cs.GR cs.CV

Etai Sella , Noam Atia , Ron Mokady , Hadar Averbuch-Elor This is my paper · ORCID
classification cs.GRcs.CV
keywords text-guidedshapeeditingpointclouddiffusioninpaintingcoordinateblendinglocalized3DTalkinversion-freePoint-E
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

BlendedPC reframes localized text-guided point cloud editing as a semantic inpainting problem: given a shape, a text prompt, and a binary mask marking the region to change, a fine-tuned diffusion model fills in the masked part under the prompt while a new inference-time coordinate blending routine keeps everything outside the edit faithful to the original. The paper's central claim is that this combination beats existing text-guided shape editing methods, ChangeIt3D and Spice-E, on every reported metric over both the full ShapeTalk test set and a newly extracted localized subset, with the largest gains in identity preservation and structural quality. The authors argue that the key to the improvement is inversion-free coordinate blending, which mixes a full-shape reconstruction stream with a masked-inpainting stream at each denoising step, avoiding the costly and often inaccurate inversion used by prior approaches.

What carries the argument

The central object is the coordinate blending operation applied to two denoised point-cloud states at every step after a transition timestep: $\hat{x}_{t-1} \leftarrow \hat{x}_{t-1} \odot M + \hat{x}_{\mathrm{recon},t-1} \odot (1 - M)$, where $M$ is the binary edit mask, $\hat{x}$ comes from the inpainting stream conditioned on the masked shape and the full text prompt, and $\hat{x}_{\mathrm{recon}}$ comes from a reconstruction stream conditioned on the full shape and an empty prompt. The operation works because the fine-tuned Inpaint-E model is trained with a permutation-variant MSE loss, so point indices stay semantically aligned during denoising and the mask computed on the input remains valid for intermediate states. Training also mixes in full-shape reconstruction samples in 10\% of iterations so the model can generate the reconstruction stream it needs at inference.

What would settle it

Measure point-index consistency directly: run the reconstruction branch on test shapes and track how well the segmentation label of each point index matches between the input and the intermediate denoised states at various timesteps. If the labels drift substantially before the transition step, or if retraining the same architecture with a permutation-invariant loss such as Chamfer distance sharply degrades the reported identity-preservation metrics, the load-bearing premise would be refuted.

Watch

Extended reading notes

Core claim

The discovery is that a point cloud diffusion model fine-tuned to inpaint masked regions from a text prompt can become a precise local editor if, during inference, the denoising trajectory is blended coordinate-by-coordinate with a simultaneous reconstruction trajectory. The reconstruction branch, conditioned on the full point cloud and an empty prompt, produces progressively denoised versions of the input; after a transition step of 20 out of 64 total steps, the inpainting branch takes over inside the mask while the reconstruction branch continues to supply the unmasked coordinates. Because the same model generates both streams, no inversion is needed, and the paper reports that the edited region also retains much of the original part's identity.

Load-bearing premise

The coordinate blending algorithm assumes that point indices keep their semantic meaning as noise is removed during denoising, so the edit mask computed on the input remains valid for intermediate reconstructed and inpainted shapes; the paper supports this with a qualitative visualization rather than a quantitative measurement.

Editorial extensions

If this is right

  • For the categories it was trained on, localized editing no longer requires a separate inversion step: the same fine-tuned diffusion model can both reconstruct the full input and inpaint the masked region, and blending the two streams at inference yields identity-preserving edits.
  • Because the reconstruction stream is also used inside the mask from step 20 onward, the edited part retains some of the original part's geometry, which the ablations attribute to the coordinate blending rather than to the inpainting model alone.
  • The approach achieves its reported gains with a relatively small fine-tuning dataset, the localized l-ShapeTalk subset, suggesting that existing text-to-point-cloud generators can be adapted for localized editing without large-scale paired training data.
  • The method is not limited to the single-category setting: a unified model trained jointly on chairs, tables, and lamps performs comparably to per-category models in the paper's experiments.

Reading between the lines

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

  • If the same permutation-variant training is preserved, the coordinate blending routine could be ported to newer or larger point-cloud diffusion backbones and to latent or token-based point representations; the paper does not test these transfers.
  • The reconstruction-conditioning percentage is tuned globally at 10\%; per-category or per-prompt tuning of this percentage could trade identity preservation against edit flexibility differently than a single fixed value.
  • The current mask generation depends on an LLM plus a semantic segmentation model, so edits are restricted to parts that can be named and segmented; instance-level or open-vocabulary segmentation would extend the method to sub-parts and multiple simultaneous edits.
  • The l-GD metric is computed by the authors with LLM-based part extraction and PointNet segmentation, and they compare it against the ChangeIt3D implementation; adopting a single standardized localized-distance protocol would make future comparisons across methods more meaningful.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 4 minor

Summary. The paper presents BlendedPC, a method for localized text-guided editing of point clouds. It fine-tunes Point-E into an inpainting model (Inpaint-E) that takes a masked point cloud and a text prompt, and introduces an inference-time coordinate blending procedure (Algorithm 1) that reconstructs the full shape under a null prompt and blends it with the inpainting trajectory according to a binary edit mask, starting at a transition timestep tr. The method is trained and evaluated on a new l-ShapeTalk subset derived from ShapeTalk with LLM-extracted part masks. The authors report quantitative gains over ChangeIt3D and Spice-E across identity, structural, and edit-fidelity metrics, a user study, and ablations showing the importance of coordinate blending.

Significance. If the claims hold, the paper offers a simple inversion-free way to achieve localized 3D edits with strong identity preservation, and the coordinate-blending idea is a useful contribution to 3D editing. Strengths include the consistent quantitative gains in Table 1, the user study, the per-category and unified-model experiments, and the detailed implementation and dataset-release plans. The main concerns are that the central mechanism rests on an unverified index-consistency assumption and that the evaluation lacks statistical rigor for some of the reported metric differences.

major comments (4)
  1. [Section 3.2.2, Algorithm 1, and Supplementary Section 7.1] The coordinate blending operation in Algorithm 1, xhat_{t-1} <- xhat_{t-1} * M + xhat_recon,t-1 * (1-M), is valid only if point indices remain semantically aligned with the input throughout denoising. The only evidence offered is a qualitative single-sample visualization (Supplementary Figure 9) plus the claim that the permutation-variant MSE loss enforces index consistency. This assumption is load-bearing for the paper's central claim that coordinate blending enables identity-preserving localized edits, and it is not supported by a quantitative measurement. Please report, for example, the overlap between the input part mask and the corresponding points in xhat_recon,t at several timesteps, averaged over the test set and across categories. Without such evidence, the stated mechanism for the GD and l-GD gains is not established.
  2. [Tables 1, 3, and Supplementary Table 9] The hyperparameters tr=20 and the 10% reconstruction ratio appear to be selected using the evaluation metrics (Section 4.3 and Supplementary Section 7.8). If this selection was performed on the test splits, the reported gains are optimistically biased. Please clarify the validation protocol and report standard deviations or confidence intervals over test samples, together with paired significance tests for the key metrics. This matters because the CLIPSim and CLIPDir differences in Table 1 are small (e.g., 0.26 vs 0.25 and 0.99 vs 1.01), while the identity-metric differences are large; the claim of outperforming alternatives across all metrics needs statistical support.
  3. [Supplementary Section 6.2, Table 5] The l-GD metric is computed with the authors' own LLM+PointNet mask pipeline rather than the ChangeIt3D script, and the two implementations yield substantially different absolute values (e.g., for Ours, 0.78 vs 0.07 on ShapeTalk). The paper argues that the trends are consistent, but because the same pipeline is used to derive the edit masks for the method's inputs, there is a risk of favorable bias toward the proposed method. Please report l-GD with the ChangeIt3D script as the primary metric, or justify in detail why the custom implementation is preferable, and describe exactly how masks are obtained for baseline outputs.
  4. [Section 3.2.1 and Algorithm 1] The coordinate blending is applied only in the 1024-point generator stage; the final 4096-point output is produced by the unmodified Point-E upsampler, which is neither mask-conditioned nor blend-aware. The paper reports metrics on the final downsampled 4096-point outputs (Supplementary Section 6.2), but provides no evidence that the upsampler preserves the identity guarantees established at the generator level. Please quantify mask-region geometric consistency before and after upsampling, or state explicitly that the identity-preservation guarantees hold only for the generator output.
minor comments (4)
  1. [Table 7 vs Table 1] The CLIPDir metric is reported with a downward arrow in Table 1 but an upward arrow in Table 7, while the definition in Eq. (4) indicates lower is better. Please correct the arrow direction and any corresponding text.
  2. [General] There are several typographical inconsistencies, including 'Shapetalk' versus 'ShapeTalk' in tables and text, 'datasts' in the Table 7 caption, 'acheives' in Section 6.2, and 'unintuative' in Section 7.1. A careful proofread is needed.
  3. [Section 3.3] The construction of l-ShapeTalk and the evaluation protocol for the full ShapeTalk set ('selecting the most relevant part for editing from each prompt') are described only briefly. Please provide the number of samples retained per category and the explicit exclusion criteria, so that the evaluation set is reproducible.
  4. [Supplementary Section 6.2] The translation of ShapeTalk prompts into more descriptive prompts via Llama 3 is a non-trivial evaluation modification. Please provide examples of the translations and state clearly whether the same translated prompts are used for all baselines, since this affects the fairness of the edit-fidelity metrics.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: the coordinate-blending identity guarantee is stated as a constructive mechanism rather than derived from the evaluation metric.

full rationale

The paper is an empirical systems paper, not a derivation, so the definitional-circularity patterns do not apply. The central coordinate-blending step is explicitly constructive: Algorithm 1 sets unmasked output points to the reconstruction branch's output, and the paper states this guarantee directly ('we guarantee that the unedited points in the output ... are nearly identical to the original point cloud'), rather than presenting the resulting l-GD improvement as an unanticipated prediction. The method is compared against external baselines (ChangeIt3D, Spice-E) on shared test splits, with a user study and multiple ablations; the l-GD implementation is cross-checked against the ChangeIt3D script in Table 5. The self-citations (Spice-E for cross-entity attention, Vox-E as a baseline) are design and baseline choices, not load-bearing justifications of the core claim. The only notable weaknesses—index-consistency supported mainly by a qualitative figure, and hyperparameters (tr, reconstruction ratio) selected on the evaluation sets—are evidence-quality and selection-bias issues, not circular derivation, and there is no equation-level reduction of the claimed result to its inputs.

Assumptions & free parameters 3 free parameters · 5 assumptions · 0 invented entities

The central method depends on three kinds of input: hand-set inference and training ratios (tr, reconstruction ratio, text-drop probability) that are tuned on the evaluation distribution; domain assumptions about the reliability of Point-E, the ShapeTalk/l-ShapeTalk data, and the LLM plus segmentation mask pipeline; and an ad hoc assumption that point indices keep their semantic identity under an MSE loss, which is load-bearing for coordinate blending and only qualitatively supported. No new physical or architectural entities are postulated beyond the fine-tuned Inpaint-E model.

free parameters (3)
  • transition_timestep_tr = 20 of T=64
    Controls how many final diffusion steps perform coordinate blending vs pure reconstruction. The authors state it was set empirically to balance identity preservation and edit flexibility, and they ablate it on evaluation data (Section 4.3, Supplementary 7.2).
  • reconstruction_sample_ratio = 10%
    Fraction of training iterations where the masked input and text are replaced by the full shape and empty prompt, enabling reconstruction denoising. Chosen after ablations on evaluation data (Supplementary 7.8).
  • text_guidance_drop_probability = 0.5
    Probability of dropping the text prompt during training to encourage reliance on structural guidance. Hand-set, mentioned in Supplementary 6.1.
assumptions (5)
  • domain assumption Point-E is a capable text-to-point-cloud diffusion backbone whose generative prior survives fine-tuning.
    Inpaint-E makes minimal architectural changes to Point-E and expects its generative abilities to be retained (Section 3.2.1).
  • ad hoc to paper Training with a permutation-variant MSE loss keeps point indices semantically consistent during denoising.
    Coordinate blending requires that mask point indices align with semantic parts across the reconstruction and inpainting chains. Only qualitative evidence is provided (Supplementary Section 7.1, Figure 9).
  • domain assumption Llama3 and PointNet reliably extract part names and segment the correct edit region from text prompts.
    Mask quality determines edit localization during training and evaluation; samples where segmentation failed were discarded (Section 3.3, Supplementary 6.1).
  • domain assumption The l-ShapeTalk subset provides valid supervision for text-guided inpainting.
    The data pipeline filters ShapeTalk prompts by part names using an LLM, and the paper assumes the remaining prompts are localized and accurately annotated (Section 3.3).
  • domain assumption The reconstruction denoising chain with an empty prompt produces intermediate point clouds close enough to the input to preserve identity outside the mask.
    The inference algorithm relies on Inpaint-E's ability to reconstruct the full shape from random noise; this is enabled by the 10% reconstruction training ratio and verified only through aggregate metrics (Section 3.2.2, Supplementary 7.8).

how reviews work

0 comments
Cite this review

Pith. "Pith review of Blended Point Cloud Diffusion for Localized Text-guided Shape Editing." pith.science (2026). https://pith.science/paper/ABITGGX2

@misc{pith2026250715399,
  author       = {Pith},
  title        = {Pith review of: Blended Point Cloud Diffusion for Localized Text-guided Shape Editing},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ABITGGX2}},
  note         = {Machine review of arXiv:2507.15399}
}
read the original abstract

Natural language offers a highly intuitive interface for enabling localized fine-grained edits of 3D shapes. However, prior works face challenges in preserving global coherence while locally modifying the input 3D shape. In this work, we introduce an inpainting-based framework for editing shapes represented as point clouds. Our approach leverages foundation 3D diffusion models for achieving localized shape edits, adding structural guidance in the form of a partial conditional shape, ensuring that other regions correctly preserve the shape's identity. Furthermore, to encourage identity preservation also within the local edited region, we propose an inference-time coordinate blending algorithm which balances reconstruction of the full shape with inpainting at a progression of noise levels during the inference process. Our coordinate blending algorithm seamlessly blends the original shape with its edited version, enabling a fine-grained editing of 3D shapes, all while circumventing the need for computationally expensive and often inaccurate inversion. Extensive experiments show that our method outperforms alternative techniques across a wide range of metrics that evaluate both fidelity to the original shape and also adherence to the textual description.

Figures

Figures reproduced from arXiv: 2507.15399 by the authors.

Figure 1
Figure 1. Given an input point cloud and a target text prompt, our [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. An overview of Inpaint-E’s training procedure. At each iteration we sample a target shape x0 and a text prompt from l￾ShapeTalk (Section 3.3) and use Llama [13] and PartNet [35] to extract a binary mask M that indicates the part that requires editing. We then use the binary mask to produce a partial point cloud that is given to Inpaint-E as guidance in addition to the text prompt and timestep t. We then add random n… view at source ↗
Figure 3
Figure 3. Inversion-Free Inference Time Coordinate Blending. A diagram illustrating our proposed inference-time mechanism, as described in Algorithm 1 and Section 3.2.2. The red points represent points generated by Inpaint-E using the empty string and input point cloud as guidance (i.e., reconstruction denoising), and the green represent points generated by Inpaint-E using the full editing prompt and the partial point cloud x… view at source ↗
Figures from the paper (9 more)
Figure 4
Figure 4. Figure 4: Qualitative comparison. We compare our method’s outputs to those of ChangeIt3D [2] and Spice-E [47]. As illus￾trated above, our method outperforms these baselines in terms of edit fidelity, identity preservation and overall visual quality. Out￾puts for all methods are …
Figure 5
Figure 5. Figure 5: Qualitative ablation results, obtained for samples from the ShapeTalk dataset. As detailed in Section 4.3, we compare our method to Inpaint-E (i.e. removing our inference-time coordi￾nate blending algorithm) and our method when tr is set to T (i.e. inpainting is perfor…
Figure 6
Figure 6. Figure 6: Image Editing Followed by Single-View 3D Reconstruction. We first use InstructPix2Pix to edit the rendered image and then apply One-2-3-45++ for single-view reconstruction. In addition to InstructPix2Pix’s challenges with precise localization, One-2-3-45++ introduces m…
Figure 7
Figure 7. Figure 7: Ablation Study for tr Values. Higher tr values increase the number of coordinate blending steps, providing better editing freedom but at the cost of inferior identity preservation. Conversely, lower tr values improve identity preservation while reducing the model’s abi…
Figure 8
Figure 8. Figure 8: Qualitative comparison. We compare our method’s outputs to those of the image editing method InstructPix2Pix [6] as well as the score distillation sampling optimization based 3D editing works Fantasia 3D [8] and Vox-E [46]. As illustrated above, our method outperforms …
Figure 9
Figure 9. Figure 9: An illustration demonstrating that the masking module is directly applicable to the intermediate denoised point clouds; additional [PITH_FULL_IMAGE:figures/full_fig_p016_9.png]
Figure 10
Figure 10. Figure 10: Results Gallery. Above we show results over various object categories including chair, table, lamp, airplane, cap, guitar, skateboard and knife [PITH_FULL_IMAGE:figures/full_fig_p017_10.png]
Figure 11
Figure 11. Figure 11: Qualitative comparison against SDFusion [10] un￾conditional part completion. We compare our method’s outputs against the unconditional part completion results of the SDFusion [10] baseline. In each row we task the methods with completing (in SDFusion’s case) or editin…
Figure 12
Figure 12. Figure 12: Limitations. As illustrated above, our approach cannot add new parts (such as arms on the left) or perform global edits that involve multiple parts (such as turning the chair into a royal chair on the right). 7.8. Ablating the proportion of reconstruction training sam…

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 2 Pith papers

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

  1. VoxHammer: Training-Free Precise and Coherent 3D Editing in Native 3D Space

    cs.CV 2025-08 conditional novelty 7.0 of 10

    A training-free 3D editing method that inverts a source asset into TRELLIS latent space and replaces latents plus attention K/V tokens in unedited regions during re-denosing.

  2. EditFlow3D: Automated Local Editing of 3D Assets with Trajectory Preservation

    cs.CV 2026-08 conditional novelty 6.0 of 10

    Mask-guided differential flow with a soft preservation loss enables training-free local 3D editing that keeps unedited regions close to the source asset.

Reference graph

Works this paper leans on

68 extracted references · 50 canonical work pages · cited by 2 Pith papers

  1. [1]

    Changeit3d: Languageas- sisted 3d shape edits and deformations

    Panos Achlioptas, Ian Huang, Minhyuk Sung, Sergey Tulyakov, and Leonidas Guibas. Changeit3d: Languageas- sisted 3d shape edits and deformations. In Conference on Computer Vision and Pattern Recognition (CVPR) , page 6,

    work page

  2. [2]

    Shapetalk: A language dataset and framework for 3d shape edits and deformations

    Panos Achlioptas, Ian Huang, Minhyuk Sung, Sergey Tulyakov, and Leonidas Guibas. Shapetalk: A language dataset and framework for 3d shape edits and deformations. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 12685–12694, 2023. 1, 2, 5, 7

    work page 2023

  3. [3]

    Blended diffusion for text-driven editing of natural images

    Omri Avrahami, Dani Lischinski, and Ohad Fried. Blended diffusion for text-driven editing of natural images. In Pro- ceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 18208–18218, 2022. 1, 3

    work page 2022

  4. [4]

    Blended latent diffusion

    Omri Avrahami, Ohad Fried, and Dani Lischinski. Blended latent diffusion. ACM transactions on graphics (TOG) , 42 (4):1–11, 2023. 1, 3

    work page 2023

  5. [5]

    Making Them Move: Mechanics, Control & Animation of Articulated Figures

    Norman Badler, Brian Barsky, and David Zeltzer. Making Them Move: Mechanics, Control & Animation of Articulated Figures. Routledge, 1990. 2

    work page 1990

  6. [6]

    In- structpix2pix: Learning to follow image editing instructions

    Tim Brooks, Aleksander Holynski, and Alexei A Efros. In- structpix2pix: Learning to follow image editing instructions. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 18392–18402, 2023. 15

    work page 2023

  7. [7]

    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. 5

    arXiv 2015

  8. [8]

    Fan- tasia3d: Disentangling geometry and appearance for high- quality text-to-3d content creation

    Rui Chen, Yongwei Chen, Ningxin Jiao, and Kui Jia. Fan- tasia3d: Disentangling geometry and appearance for high- quality text-to-3d content creation. In Proceedings of the IEEE/CVF international conference on computer vision , pages 22246–22256, 2023. 1, 8, 14, 15

    work page 2023

Show all 68 references
  1. [9]

    Tango: Text-driven photorealistic and robust 3d styliza- tion via lighting decomposition

    Yongwei Chen, Rui Chen, Jiabao Lei, Yabin Zhang, and Kui Jia. Tango: Text-driven photorealistic and robust 3d styliza- tion via lighting decomposition. Advances in Neural Infor- mation Processing Systems, 35:30923–30936, 2022. 2

    2022

  2. [10]

    Sdfusion: Multimodal 3d shape completion, reconstruction, and generation

    Yen-Chi Cheng, Hsin-Ying Lee, Sergey Tulyakov, Alexan- der G Schwing, and Liang-Yan Gui. Sdfusion: Multimodal 3d shape completion, reconstruction, and generation. InPro- ceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 4456–4465, 2023. 3, 16, 18

    2023

  3. [11]

    Diffcomplete: Diffusion-based generative 3d shape completion

    Ruihang Chu, Enze Xie, Shentong Mo, Zhenguo Li, Matthias Nießner, Chi-Wing Fu, and Jiaya Jia. Diffcomplete: Diffusion-based generative 3d shape completion. Advances in Neural Information Processing Systems, 36, 2024. 2, 3

    2024

  4. [12]

    Diffusion models beat gans on image synthesis

    Prafulla Dhariwal and Alexander Nichol. Diffusion models beat gans on image synthesis. Advances in neural informa- tion processing systems, 34:8780–8794, 2021. 3

    2021

  5. [13]

    The llama 3 herd of models

    Abhimanyu Dubey, Abhinav Jauhri, Abhinav Pandey, Ab- hishek Kadian, Ahmad Al-Dahle, Aiesha Letman, Akhil Mathur, Alan Schelten, Amy Yang, Angela Fan, et al. The llama 3 herd of models. arXiv preprint arXiv:2407.21783 ,

    arXiv

  6. [14]

    Sc-diff: 3d shape completion with latent diffu- sion models

    Juan D Galvis, Xingxing Zuo, Simon Schaefer, and Stefan Leutengger. Sc-diff: 3d shape completion with latent diffu- sion models. arXiv preprint arXiv:2403.12470, 2024. 3

    2024 arXiv

  7. [15]

    Textdeformer: Geometry manipu- lation using text guidance

    William Gao, Noam Aigerman, Thibault Groueix, V ova Kim, and Rana Hanocka. Textdeformer: Geometry manipu- lation using text guidance. In ACM SIGGRAPH 2023 Con- ference Proceedings, pages 1–11, 2023. 2

    2023

  8. [16]

    Generative adversarial networks

    Ian Goodfellow, Jean Pouget-Abadie, Mehdi Mirza, Bing Xu, David Warde-Farley, Sherjil Ozair, Aaron Courville, and Yoshua Bengio. Generative adversarial networks. Commu- nications of the ACM, 63(11):139–144, 2020. 3

    2020

  9. [17]

    Dualsdf: Semantic shape manipulation using a two-level representation

    Zekun Hao, Hadar Averbuch-Elor, Noah Snavely, and Serge Belongie. Dualsdf: Semantic shape manipulation using a two-level representation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages 7631–7641, 2020. 2

    2020

  10. [18]

    Prompt-to-prompt im- age editing with cross attention control

    Amir Hertz, Ron Mokady, Jay Tenenbaum, Kfir Aberman, Yael Pritch, and Daniel Cohen-Or. Prompt-to-prompt im- age editing with cross attention control. arXiv preprint arXiv:2208.01626, 2022. 5

    2022 arXiv

  11. [19]

    Spaghetti: Editing implicit shapes through part aware generation.ACM Transactions on Graph- ics (TOG), 41(4):1–20, 2022

    Amir Hertz, Or Perel, Raja Giryes, Olga Sorkine-Hornung, and Daniel Cohen-Or. Spaghetti: Editing implicit shapes through part aware generation.ACM Transactions on Graph- ics (TOG), 41(4):1–20, 2022. 2

    2022

  12. [20]

    Gans trained by a two time-scale update rule converge to a local nash equilib- rium

    Martin Heusel, Hubert Ramsauer, Thomas Unterthiner, Bernhard Nessler, and Sepp Hochreiter. Gans trained by a two time-scale update rule converge to a local nash equilib- rium. Advances in neural information processing systems , 30, 2017. 6

    2017

  13. [21]

    Denoising dif- fusion probabilistic models

    Jonathan Ho, Ajay Jain, and Pieter Abbeel. Denoising dif- fusion probabilistic models. Advances in neural information processing systems, 33:6840–6851, 2020. 3

    2020

  14. [22]

    Cns-edit: 3d shape editing via coupled neural shape optimization

    Jingyu Hu, Ka-Hei Hui, Zhengzhe Liu, Hao Zhang, and Chi- Wing Fu. Cns-edit: 3d shape editing via coupled neural shape optimization. In ACM SIGGRAPH 2024 Conference Papers, pages 1–12, 2024. 2

    2024

  15. [23]

    Ladis: Lan- guage disentanglement for 3d shape editing

    Ian Huang, Panos Achlioptas, Tianyi Zhang, Sergey Tulyakov, Minhyuk Sung, and Leonidas Guibas. Ladis: Lan- guage disentanglement for 3d shape editing. arXiv preprint arXiv:2212.05011, 2022. 2

    2022 arXiv

  16. [24]

    Zero-shot text-guided object genera- tion with dream fields

    Ajay Jain, Ben Mildenhall, Jonathan T Barron, Pieter Abbeel, and Ben Poole. Zero-shot text-guided object genera- tion with dream fields. In Proceedings of the IEEE/CVF con- ference on computer vision and pattern recognition , pages 867–876, 2022. 2

    2022

  17. [25]

    Salad: Part-level latent diffusion for 3d shape gen- eration and manipulation

    Juil Koo, Seungwoo Yoo, Minh Hieu Nguyen, and Minhyuk Sung. Salad: Part-level latent diffusion for 3d shape gen- eration and manipulation. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 14441– 14451, 2023. 3

    2023

  18. [26]

    Understanding pure clip guidance for voxel grid nerf models

    Han-Hung Lee and Angel X Chang. Understanding pure clip guidance for voxel grid nerf models. arXiv preprint arXiv:2209.15172, 2022. 2

    2022 arXiv

  19. [27]

    ishapediting: In- telligent shape editing with diffusion models

    Jing Li, Juyong Zhang, and Falai Chen. ishapediting: In- telligent shape editing with diffusion models. In Computer Graphics Forum, page e15253. Wiley Online Library, 2024. 2

    2024

  20. [28]

    Magic3d: High-resolution text-to-3d content creation

    Chen-Hsuan Lin, Jun Gao, Luming Tang, Towaki Takikawa, Xiaohui Zeng, Xun Huang, Karsten Kreis, Sanja Fidler, Ming-Yu Liu, and Tsung-Yi Lin. Magic3d: High-resolution text-to-3d content creation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognitio...

    2023

  21. [29]

    One-2-3-45++: Fast single im- age to 3d objects with consistent multi-view generation and 3d diffusion

    Minghua Liu, Ruoxi Shi, Linghao Chen, Zhuoyang Zhang, Chao Xu, Xinyue Wei, Hansheng Chen, Chong Zeng, Ji- ayuan Gu, and Hao Su. One-2-3-45++: Fast single im- age to 3d objects with consistent multi-view generation and 3d diffusion. In Proceedings of the IEEE/CVF Conference on ...

    2024

  22. [30]

    Exim: A hybrid explicit- implicit representation for text-guided 3d shape generation

    Zhengzhe Liu, Jingyu Hu, Ka-Hei Hui, Xiaojuan Qi, Daniel Cohen-Or, and Chi-Wing Fu. Exim: A hybrid explicit- implicit representation for text-guided 3d shape generation. ACM Transactions on Graphics (TOG), 42(6):1–12, 2023. 2

    2023

  23. [31]

    Repaint: Inpainting using denoising diffusion probabilistic models

    Andreas Lugmayr, Martin Danelljan, Andres Romero, Fisher Yu, Radu Timofte, and Luc Van Gool. Repaint: Inpainting using denoising diffusion probabilistic models. In Proceed- ings of the IEEE/CVF conference on computer vision and pattern recognition, pages 11461–11471, 2022. 3, 16

    2022

  24. [32]

    Joint-dependent local deformations for hand anima- tion and object grasping

    Thalmann Magnenat, Richard Laperri `ere, and Daniel Thal- mann. Joint-dependent local deformations for hand anima- tion and object grasping. In Proceedings of Graphics Inter- face’88, pages 26–33. Canadian Inf. Process. Soc, 1988. 2

    1988

  25. [33]

    Text2mesh: Text-driven neural stylization for meshes

    Oscar Michel, Roi Bar-On, Richard Liu, Sagie Benaim, and Rana Hanocka. Text2mesh: Text-driven neural stylization for meshes. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 13492– 13502, 2022. 2

    2022

  26. [34]

    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

  27. [35]

    Partnet: A large- scale benchmark for fine-grained and hierarchical part-level 3d object understanding

    Kaichun Mo, Shilin Zhu, Angel X Chang, Li Yi, Subarna Tripathi, Leonidas J Guibas, and Hao Su. Partnet: A large- scale benchmark for fine-grained and hierarchical part-level 3d object understanding. In Proceedings of the IEEE/CVF conference on computer vision and pattern recog...

    2019

  28. [36]

    Null-text inversion for editing real im- ages using guided diffusion models

    Ron Mokady, Amir Hertz, Kfir Aberman, Yael Pritch, and Daniel Cohen-Or. Null-text inversion for editing real im- ages using guided diffusion models. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 6038–6047, 2023. 5

    2023

  29. [37]

    Glide: Towards photorealistic image generation and editing with text-guided diffusion models.arXiv preprint arXiv:2112.10741, 2021

    Alex Nichol, Prafulla Dhariwal, Aditya Ramesh, Pranav Shyam, Pamela Mishkin, Bob McGrew, Ilya Sutskever, and Mark Chen. Glide: Towards photorealistic image generation and editing with text-guided diffusion models.arXiv preprint arXiv:2112.10741, 2021. 3

    2021 arXiv

  30. [38]

    Point-e: A system for generat- ing 3d point clouds from complex prompts

    Alex Nichol, Heewoo Jun, Prafulla Dhariwal, Pamela Mishkin, and Mark Chen. Point-e: A system for generat- ing 3d point clouds from complex prompts. arXiv preprint arXiv:2212.08751, 2022. 1, 4

    2022 arXiv

  31. [39]

    Lazy diffusion transformer for interactive image editing

    Yotam Nitzan, Zongze Wu, Richard Zhang, Eli Shechtman, Daniel Cohen-Or, Taesung Park, and Micha¨el Gharbi. Lazy diffusion transformer for interactive image editing. arXiv preprint arXiv:2404.12382, 2024. 5

    2024 arXiv

  32. [40]

    Gsedit: Efficient text-guided edit- ing of 3d objects via gaussian splatting

    Francesco Palandra, Andrea Sanchietti, Daniele Baieri, and Emanuele Rodol `a. Gsedit: Efficient text-guided edit- ing of 3d objects via gaussian splatting. arXiv preprint arXiv:2403.05154, 2024. 2

    2024 arXiv

  33. [41]

    Geocode: Interpretable shape programs

    Ofek Pearl, Itai Lang, Yuhua Hu, Raymond A Yeh, and Rana Hanocka. Geocode: Interpretable shape programs. arXiv preprint arXiv:2212.11715, 2022. 2

    2022 arXiv

  34. [42]

    Dreamfusion: Text-to-3d using 2d diffusion

    Ben Poole, Ajay Jain, Jonathan T Barron, and Ben Milden- hall. Dreamfusion: Text-to-3d using 2d diffusion. arXiv preprint arXiv:2209.14988, 2022. 1, 2, 14

    2022 arXiv

  35. [43]

    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,

  36. [44]

    Learning transferable visual models from natural language supervi- sion

    Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, et al. Learning transferable visual models from natural language supervi- sion. In International conference on machine learning, ...

    2021

  37. [45]

    High-resolution image synthesis with latent diffusion models

    Robin Rombach, Andreas Blattmann, Dominik Lorenz, Patrick Esser, and Bj ¨orn Ommer. High-resolution image synthesis with latent diffusion models. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 10684–10695, 2022. 1, 3, 5

    2022

  38. [46]

    V ox-e: Text-guided voxel editing of 3d ob- jects

    Etai Sella, Gal Fiebelman, Peter Hedman, and Hadar Averbuch-Elor. V ox-e: Text-guided voxel editing of 3d ob- jects. In Proceedings of the IEEE/CVF International Con- ference on Computer Vision, pages 430–440, 2023. 1, 6, 8, 14, 15

    2023

  39. [47]

    Spice· e: Structural priors in 3d diffusion using cross- entity attention

    Etai Sella, Gal Fiebelman, Noam Atia, and Hadar Averbuch- Elor. Spice· e: Structural priors in 3d diffusion using cross- entity attention. In ACM SIGGRAPH 2024 Conference Pa- pers, pages 1–11, 2024. 1, 2, 4, 6, 7, 18

    2024

  40. [48]

    3d point cloud generative adversarial network based on tree structured graph convolutions

    Dong Wook Shu, Sung Woo Park, and Junseok Kwon. 3d point cloud generative adversarial network based on tree structured graph convolutions. In Proceedings of the IEEE/CVF international conference on computer vision , pages 3859–3868, 2019. 6

    2019

  41. [49]

    Shapewalk: Composi- tional shape editing through language-guided chains

    Habib Slim and Mohamed Elhoseiny. Shapewalk: Composi- tional shape editing through language-guided chains. InPro- ceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 22574–22583, 2024. 2

    2024

  42. [50]

    Modelnet: Towards a datacenter emulation environment

    Kashi Venkatesh Vishwanath, Diwaker Gupta, Amin Vahdat, and Ken Yocum. Modelnet: Towards a datacenter emulation environment. In 2009 IEEE Ninth International Conference on Peer-to-Peer Computing, pages 81–82. IEEE, 2009. 5

    2009

  43. [51]

    Clip-nerf: Text-and-image driven manip- ulation of neural radiance fields

    Can Wang, Menglei Chai, Mingming He, Dongdong Chen, and Jing Liao. Clip-nerf: Text-and-image driven manip- ulation of neural radiance fields. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 3835–3844, 2022. 2

    2022

  44. [52]

    Gaussianeditor: Editing 3d gaussians delicately with text instructions

    Junjie Wang, Jiemin Fang, Xiaopeng Zhang, Lingxi Xie, and Qi Tian. Gaussianeditor: Editing 3d gaussians delicately with text instructions. In Proceedings of the IEEE/CVF con- ference on computer vision and pattern recognition , pages 20902–20911, 2024. 2

    2024

  45. [53]

    Shape inpainting using 3d generative ad- versarial network and recurrent convolutional networks

    Weiyue Wang, Qiangui Huang, Suya You, Chao Yang, and Ulrich Neumann. Shape inpainting using 3d generative ad- versarial network and recurrent convolutional networks. In Proceedings of the IEEE international conference on com- puter vision, pages 2298–2306, 2017. 2, 3

    2017

  46. [54]

    Cas- caded refinement network for point cloud completion

    Xiaogang Wang, Marcelo H Ang Jr, and Gim Hee Lee. Cas- caded refinement network for point cloud completion. In Proceedings of the IEEE/CVF conference on computer vi- sion and pattern recognition, pages 790–799, 2020. 3

    2020

  47. [55]

    Gaussctrl: Multi-view consistent text-driven 3d gaussian splatting edit- ing

    Jing Wu, Jia-Wang Bian, Xinghui Li, Guangrun Wang, Ian Reid, Philip Torr, and Victor Adrian Prisacariu. Gaussctrl: Multi-view consistent text-driven 3d gaussian splatting edit- ing. In European Conference on Computer Vision, pages 55–

  48. [56]

    Snowflakenet: Point cloud completion by snowflake point deconvolution with skip-transformer

    Peng Xiang, Xin Wen, Yu-Shen Liu, Yan-Pei Cao, Pengfei Wan, Wen Zheng, and Zhizhong Han. Snowflakenet: Point cloud completion by snowflake point deconvolution with skip-transformer. In Proceedings of the IEEE/CVF inter- national conference on computer vision , pages 5499–5509,

  49. [57]

    Smartbrush: Text and shape guided object inpainting with diffusion model

    Shaoan Xie, Zhifei Zhang, Zhe Lin, Tobias Hinz, and Kun Zhang. Smartbrush: Text and shape guided object inpainting with diffusion model. In Proceedings of the IEEE/CVF Con- ference on Computer Vision and Pattern Recognition, pages 22428–22437, 2023. 3, 5

    2023

  50. [58]

    Instructp2p: Learning to edit 3d point clouds with text instructions

    Jiale Xu, Xintao Wang, Yan-Pei Cao, Weihao Cheng, Ying Shan, and Shenghua Gao. Instructp2p: Learning to edit 3d point clouds with text instructions. arXiv preprint arXiv:2306.07154, 2023. 2, 8, 13

    2023 arXiv

  51. [59]

    Shapeformer: Transformer-based shape completion via sparse representa- tion

    Xingguang Yan, Liqiang Lin, Niloy J Mitra, Dani Lischin- ski, Daniel Cohen-Or, and Hui Huang. Shapeformer: Transformer-based shape completion via sparse representa- tion. In Proceedings of the IEEE/CVF Conference on Com- puter Vision and Pattern Recognition , pages 6239–6249,

  52. [60]

    Neumesh: Learning disentangled neural mesh-based implicit field for geometry and texture editing

    Bangbang Yang, Chong Bao, Junyi Zeng, Hujun Bao, Yinda Zhang, Zhaopeng Cui, and Guofeng Zhang. Neumesh: Learning disentangled neural mesh-based implicit field for geometry and texture editing. In European Conference on Computer Vision, pages 597–614. Springer, 2022. 2

    2022

  53. [61]

    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. 2, 3

    2021

  54. [62]

    Unsupervised 3d shape completion through gan inver- sion

    Junzhe Zhang, Xinyi Chen, Zhongang Cai, Liang Pan, Haiyu Zhao, Shuai Yi, Chai Kiat Yeo, Bo Dai, and Chen Change Loy. Unsupervised 3d shape completion through gan inver- sion. In Proceedings of the IEEE/CVF Conference on Com- puter Vision and Pattern Recognition , pages 1768–1777,

  55. [63]

    Point cloud completion via skeleton- detail transformer

    Wenxiao Zhang, Huajian Zhou, Zhen Dong, Jun Liu, Qingan Yan, and Chunxia Xiao. Point cloud completion via skeleton- detail transformer. IEEE Transactions on Visualization and Computer Graphics, 29(10):4229–4242, 2022. 3

    2022

  56. [64]

    3d shape generation and completion through point-voxel diffusion

    Linqi Zhou, Yilun Du, and Jiajun Wu. 3d shape generation and completion through point-voxel diffusion. In Proceed- ings of the IEEE/CVF International Conference on Com- puter Vision (ICCV), pages 5826–5835, 2021. 3 Blended Point Cloud Diffusion for Localized Text-guided Shape ...

    2021

  57. [65]

    non-edit

    Implementation Details 6.1. Training We trained our model for each category using object- specific ShapeTalk and l-ShapeTalk subsets. To evaluate generalization, we also trained a unified model across all three categories (Chair, Table, and Lamp). For all models, we used a bat...

  58. [66]

    resampling

    Additional Experiments To better view 3D results, we recommended viewing the supplemental HTML page, which includs fly-through vi- sualizations demonstrating the quality of our results from multiple views. 7.1. Comparison to Semantic Editing Paradigms In this section, we compl...

  59. [67]

    Recon %”) our identity preservation capabilities are significantly ham- pered, yielding high GD and l-GD scores. Conversely, high “Recon %

    baseline. In each row we task the methods with completing (in SDFusion’s case) or editing* (in our case) a different part. As these results show, SDFusion’s ability to preserve identity across all regions of the shape is limited, in comparison to our approach that performs loc...

  60. [68]

    As illustrated in the figure, our inpainting-based approach restricts the ability to generate entirely new objects or parts

    Limitations While our method performs well in most scenarios, it has certain limitations; see Figure 12. As illustrated in the figure, our inpainting-based approach restricts the ability to generate entirely new objects or parts. BlendedPC is specifically designed for localize...

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.