Pith. sign in

REVIEW 5 major objections 4 minor 56 references

SCULPT: Subtractive Composition for 3D Part Generation

T0 review · 5 major / 4 minor · reviewed 2026-08-14 · deepseek-v4-flash

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

Pith's one-line read Part-aware 3D generation can be done by subtractive composition, repeatedly splitting a part and the remainder from a whole-object latent, and the paper reports the best geometry on PartObjaverse.

desk verdict The recurrent part-remainder formulation is genuinely new and well demonstrated, but the state-of-the-art claim overreaches: the closest baselines are missing and the margins over the strongest included one are small and unquantified. read the letter →

arxiv 2608.13541 v1 pith:VEHPUDP4 submitted 2026-08-13 cs.CV cs.GR

Sikuang Li , Chen Yang , Jiemin Fang , Jiazhong Cen , Yuhe Wei , Jichen Pang , Wei Shen , Qi Tian This is my paper · ORCID
classification cs.CVcs.GR
keywords 3Dpartgenerationsubtractivecompositionpart-awarestructuredlatentrectifiedflowtransformersrecurrentdecompositionimage-to-3Dsegmentationandassembly
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

SCULPT claims that part-aware 3D generation is best done subtractively: instead of labeling an already generated shape or synthesizing parts and then reconciling them, the model starts from a complete object in a generative 3D latent space and repeatedly predicts one extracted part together with the object that remains. Each split is a fixed-signature operation, so the number of parts is determined by how many splits the rollout performs, up to a safety cap, rather than by a fixed number of output slots. The part and remainder are denoised together on overlapping supports in a shared object frame, which is designed to keep contact boundaries, geometry, and materials coherent. On the part-annotated PartObjaverse benchmark the paper reports the lowest Chamfer distance and best F1 at the part, semantic-group, and object levels, meaning the parts are individually accurate and the assembled whole stays faithful. If the claim holds, part editing, material assignment, and reassembly become outputs of generation rather than post-hoc fixes.

What carries the argument

The load-bearing object is the joint split predictor, a recurrent two-output module that turns the current 3D state into one extracted part and the updated remainder. It operates on the structured latent of the backbone generator, which has sparse-structure, geometry, and material stages in a shared $[-1,1]^3$ object frame; part and remainder are packed onto the union of their sparse supports, denoised jointly by rectified-flow transformer blocks initialized from the pretrained whole-object generator, and controlled by a remainder-injection branch whose residual projections start at zero. A composition loss compares the differentiable union of the predicted part and remainder occupancies with the incoming remainder's support, allowing an overlapping interface shell instead of a disjoint voxel partition, and inference-time clipping keeps every predicted support inside the current object. This machinery is what lets one fixed-signature operation express an object-dependent number of parts through rollout length, and it is what the paper credits for the part-level and object-level geometry scores.

What would settle it

Take a held-out object, encode its complete mesh and also encode one of its part meshes together with the complementary remainder mesh in the backbone's structured latent, decode each independently, and merge the two outputs; if the merged Chamfer distance to the original complete mesh is much larger than the reported object-level value of about 0.0020, or if visible gaps appear at the shared interface, the subtractive-composition assumption fails. A cheaper check is to compare the geometry and material colors on the overlapping interface-shell voxels predicted from the part side and from the remainder side of a split and measure their disagreement.

Watch

Extended reading notes

Core claim

The paper's central discovery is that a variable-cardinality part decomposition can be produced by composing a single two-output prediction rather than by predicting the whole part set at once. Given a complete-object latent from a pretrained image-to-3D generator, a joint split predictor emits the next part and the updated remainder in a coupled denoising trajectory; the predictor is then applied again to the remainder, and the rollout stops when the remainder's support is empty or a cap of 24 splits is reached. Because the part and remainder stay in the coordinate frame of the original object and may share overlapping boundary voxels, the boundary between them is decided during generation instead of imposed by a voxel-face partition. The paper reports that on PartObjaverse this yields the best Chamfer distance and best F1 at threshold 0.05 at the part, semantic-group, and object levels, with object-level Chamfer distance 0.0020 and F1 at threshold 0.05 of 0.9212, and that the approach also produces textured decompositions from text-to-image inputs and a real photograph. The authors would summarize the result as: subtractive composition turns the difficulty of modeling how many parts an object has into the easier question of when the remaining object is empty.

Load-bearing premise

The load-bearing premise is that the pretrained generator's internal 3D representation can faithfully represent partial sub-objects and overlapping contact regions, so a part and a remainder decoded separately and merged reproduce the complete object.

Editorial extensions

If this is right

  • Part structure becomes a native output of generation: the extracted parts carry geometry, material, and texture and can be edited or reassigned without a separate segmentation step.
  • Variable part counts emerge from rollout termination, so the model is not limited by a predefined number of slots or an image-space layout plan.
  • Because every part is decoded in the shared object frame, the assembled union needs no registration, rescaling, or snapping, which preserves the complete-object silhouette.
  • The benchmark numbers imply that fine-grained decomposition does not trade away whole-object fidelity: the best object-level scores are obtained by the same model that produces the best parts.
  • The same rollout extends beyond the benchmark to text-to-image inputs and a real photograph, suggesting the subtractive procedure transfers to casual images.

Reading between the lines

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

  • A natural extension the paper does not pursue is porting the split interface to other whole-object generators that have a sparse structured latent and a decoder, since the formulation itself does not depend on the specific backbone.
  • The deterministic centroid-based extraction order is a convention rather than a requirement; a learned or user-selected split order could expose hierarchical or part-graph structure using the same machinery.
  • The overlapping interface shell makes a direct boundary-coherence test possible: comparing geometry and material on voxels predicted from both sides of a split would quantify how much of the gain comes from joint denoising.
  • The paper's cap of 24 splits is a practical bound, and the demonstrated recursive application to already-extracted parts suggests that a calibrated stopping rule could make decomposition depth-adaptive rather than capped.
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

5 major / 4 minor

Summary. The paper introduces SCULPT, a recurrent subtractive decomposition method for part-aware 3D generation. Starting from a complete-object latent produced by TRELLIS.2 from a conditioning image, a joint split predictor iteratively predicts one extracted part together with the updated remainder in the shared object frame, using overlapping sparse supports, a composition loss, and inference-time support clipping. The rollout terminates when the remainder support is empty or when a fixed cap Kmax=24 is reached. Training uses part-annotated PartVerse-XL, and evaluation on PartObjaverse reports Chamfer distance and F1 at the part, semantic-group, and object levels, together with ablations and qualitative examples. The paper claims state-of-the-art geometry on PartObjaverse and presents the subtractive formulation as the main contribution.

Significance. The subtractive formulation is a genuinely different design point from additive generation and post-hoc segmentation, and the paper's ablations support the importance of full-stage adaptation, composition supervision, and inference-time clipping. If the empirical result holds over a complete baseline set, SCULPT would provide a practical way to obtain aligned, textured parts while preserving whole-object quality. Strengths include a clear methodological presentation, use of a strong pretrained backbone with zero-initialized control injection, and a three-level evaluation that separates component fidelity from assembly fidelity. The main weakness is that the state-of-the-art claim is currently supported only against a subset of the field, with small margins and no uncertainty quantification.

major comments (5)
  1. [Section 2.3 / Table 1] The related work identifies UniPart and AutoPartGen as the closest part-structured generation methods to SCULPT, yet neither appears in Table 1; the comparison is limited to Part123, OmniPart, and segmentation-plus-completion pipelines. Since the headline claim is state-of-the-art geometry on PartObjaverse, the absence of the two closest baselines leaves that claim unsupported. Please add these methods under the same evaluation protocol or explicitly restrict the claim to the evaluated baseline set.
  2. [Section 4.2 / Table 1] The margin over the strongest included baseline, TRELLIS.2+PartField, is small (part-level CD 0.0107 versus 0.0115; object-level CD 0.0020 versus 0.0021), and SCULPT is numerically worse on part-level and semantic-group F1@.1 (0.8858 versus 0.8897 and 0.8851 versus 0.8903). No variance, confidence intervals, or paired significance tests are reported, so the reader cannot tell whether the ordering is stable. Please report standard deviations or confidence intervals for all metrics, per-class results for all baselines, and a paired test over the 200 evaluation meshes.
  3. [Section 4.1 / Metrics] The Hungarian matching protocol for the part-level scores is underspecified for variable cardinailty: the paper does not state the cost matrix used for matching or how unmatched predicted and reference parts are penalized. The semantic-group procedure, in which each matched prediction inherits the semantic label of its reference part, can also conceal over-segmentation errors if extra predictions are matched incorrectly. Please specify the matching and scoring rules in enough detail that all methods are compared under exactly the same protocol.
  4. [Section 3.2 / 3.3] The method assumes that the structured latent and decoder can faithfully represent arbitrary sub-objects and overlapping interface shells, since parts and remainders are encoded and decoded independently in the shared object frame. This assumption is never verified. Please add a reconstruction experiment in which annotated part and remainder meshes are encoded, decoded, and compared against the original meshes with CD/F1 and qualitative inspection of the overlap boundaries. If reconstruction degrades for partial or overlapping supports, the central geometry claim needs qualification.
  5. [Section 3.5 / Algorithm 1] The paper does not report how often the rollout hits the cap Kmax=24 on PartObjaverse or the distribution of predicted part counts. If the cap is frequently reached, the claim that the number of parts follows the rollout length is weakened, and the final retained remainder becomes a catch-all output. Please report the cap hit rate and predicted part-count statistics, and compare them with the ground-truth part-count distribution on the evaluation set.
minor comments (4)
  1. [Abstract / Section 4.2] The abstract claims 'state-of-the-art geometry' without qualification, but Table 1 shows that SCULPT is not the best on part-level and semantic-group F1@.1; please qualify the claim to match the reported metrics.
  2. [Section 3.2] The term 'O-V oxel' appears to be a typo or needs a consistent definition; the manuscript uses 'O-Voxel' elsewhere in the same section.
  3. [Table 4] Table 4 uses the headers 'F1-0.1' and 'F1-0.05' while Table 1 uses 'F1@.1' and 'F1@.05'; please unify the notation.
  4. [Section 4.3] The qualitative section does not identify the text-to-image model used to produce the fifth row of Figure 4; please state the model for reproducibility.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: SCULPT's central claims rest on external-benchmark evaluation with a stated dedup protocol.

full rationale

The derivation chain is self-contained. The joint split predictor Pθ is trained on ground-truth part–remainder tuples from PartVerse-XL (Eqs. 6–9); its targets are encoded from annotated part meshes, not from its own outputs, so the flow-matching and composition losses (Eqs. 16–20) cannot be minimized by copying the input remainder. Inference (Algorithm 1) is a recurrent rollout of the same predictor, and no parameter is fitted to PartObjaverse. The quantitative claims are comparisons on an external benchmark (Table 1) with a stated SHA-256 dedup step between training and evaluation (Sec. 4.1, App. B.2), so the SOTA claim is not defined in terms of SCULPT's own outputs. The self-citations in related work (Cen et al. 2026; Wu et al. 2025; Yang et al. 2024a) are contextual and not used to justify any prediction. The absence of UniPart and AutoPartGen from Table 1 and the small, unquantified margins are scope-of-evidence concerns, not circularity. No equation reduces a predicted quantity to an input by construction.

Assumptions & free parameters 5 free parameters · 5 assumptions · 1 invented entities

The central claim rests on the pretrained TRELLIS.2 latent space, the quality of PartVerse-XL annotations, and several hand-chosen hyperparameters (lambda_comp, K_max, tau_occ, token filter, extraction order). No new physical entities are introduced; the interface shell is a design artifact, not an independently evidenced entity.

free parameters (5)
  • Composition loss weight lambda_comp = 0.1
    Chosen hyperparameter in Eq. 20; the ablation in Table 2 shows removing it changes part-level CD from 0.0107 to 0.0279, so the central results depend on this choice.
  • Rollout cap K_max = 24
    Safety cap in Algorithm 1; limits the maximum number of parts to 24, so objects requiring more parts need recursive application. It is a fixed hyperparameter, not fitted to the benchmark.
  • Occupancy threshold tau_occ = not stated
    Load-bearing threshold in Eq. 21 and Algorithm 1 for support clipping and termination; its value is never given in the paper or appendix, hindering exact reproduction.
  • Token-count split filter for geometry and material stages = 8 to 8192 tokens
    Splits outside this range are omitted from geometry and material training (Section 4.1), removing 1.5% of splits and 2.3% of assets; this hand-chosen selection may bias the learned distribution.
  • Lexicographic extraction order (z then x then y centroids) = z, then x, then y
    Deterministic training order defined in Section 3.3 that creates the part-remainder supervision sequences; the model learns to extract in this order, and the order is a hand-chosen design decision.
assumptions (5)
  • domain assumption TRELLIS.2 structured latents can represent arbitrary part and remainder sub-objects in the shared object frame, including overlapping boundary voxels.
    The entire recurrent split (Eqs. 1-2, Section 3.2) assumes that a part latent and a remainder latent decoded independently can be unioned to reproduce the object; this is not proven and is only indirectly tested via object-level metrics.
  • domain assumption PartVerse-XL (FullPart) annotations are consistent, human-refined, and sufficiently complete to supervise 330,455 part-remainder splits.
    Section 4.1 relies on this dataset; if annotations are noisy or inconsistent, the split predictor learns noise, and no annotation quality analysis is provided.
  • domain assumption The SHA-256-based dedup fully prevents evaluation leakage on PartObjaverse.
    Sections 4.1 and B.2: exact-match removal does not remove near-duplicate objects, which are common in Objaverse; residual leakage could inflate benchmark scores.
  • standard math Rectified flow matching (Eqs. 4-5) is a valid generative training objective for the packed part-remainder target.
    Standard flow-matching theory; the paper inherits it from the backbone, and no re-derivation is needed.
  • ad hoc to paper The deterministic z-x-y centroid ordering induces a learnable part extraction sequence.
    Section 3.3: this ordering is introduced specifically for this paper, is not grounded in semantic criteria, and may not match how parts are conceptually organized.
invented entities (1)
  • Interface shell (overlapping support region)
    purpose: Allows boundary voxels to belong to both the extracted part and the remainder, avoiding hard voxel-face partitions at part boundaries.
    A modeling construct introduced in Sections 3.3-3.4; its benefit is shown only through end-to-end ablations, and there is no external, falsifiable handle measuring the interface shell itself.

how reviews work

0 comments
Cite this review

Pith. "Pith review of SCULPT: Subtractive Composition for 3D Part Generation." pith.science (2026). https://pith.science/paper/VEHPUDP4

@misc{pith2026260813541,
  author       = {Pith},
  title        = {Pith review of: SCULPT: Subtractive Composition for 3D Part Generation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/VEHPUDP4}},
  note         = {Machine review of arXiv:2608.13541}
}
read the original abstract

Part-aware 3D generation aims to create digital assets that are coherent as complete objects while exposing structural parts for editing, material assignment, animation, and reuse. Existing methods impose this structure outside the native generation loop: segmentation-based methods partition an already generated shape, while additive methods synthesize parts from predefined layouts, boxes, or tokens and then reconcile them into a whole. The former preserves the generated geometry but fixes the object before part boundaries are determined; the latter exposes part cardinality but often leaves shared boundaries vulnerable to gaps, interpenetrations, and material discontinuities. In this paper, we propose SCULPT, a framework that addresses these challenges through subtractive composition. Given a complete object represented in a structured 3D latent space, SCULPT iteratively applies a joint split predictor to generate one extracted part together with the remaining object. The predictor performs a coupled denoising process conditioned on both the image and the current 3D state, so the extracted part and updated remainder are generated together rather than reconciled after generation. The joint split predictor processes both outputs on the union of their native sparse 3D supports, allowing neighboring supports to overlap rather than imposing a disjoint voxel partition. The rollout ends when the remainder support becomes empty or reaches a fixed safety cap, allowing the number of generated parts to adapt to each object within that bound. Extensive experiments demonstrate state-of-the-art geometry on PartObjaverse while preserving strong complete-object reconstruction after part assembly. Results on four dataset images, one text-to-image-generated input, and one real-world photograph further show fine-grained textured part decomposition beyond the benchmark.

Figures

Figures reproduced from arXiv: 2608.13541 by the authors.

Figure 1
Figure 1. SCULPT generates complete 3D objects together with coherent part structure. Starting from an image-conditioned complete object, SCULPT subtractively extracts semantically meaningful components while retaining their alignment in the original object frame. (a) Part generation: complete objects and their extracted parts across diverse object categories, with coherent geometry, boundaries, and appearance. (b) In-the-wil… view at source ↗
Figure 2
Figure 2. SCULPT turns whole-object generation into recurrent part–remainder prediction. We first use TRELLIS. 2 mapping the conditioning image I to a complete structured latent z0. At split i, the image- and remainder-conditioned predictor Pθ jointly denoises an extracted part zˆi (blue) and the updated remainder zi (red); only the remainder state is carried into the next split, until its sparse support becomes empty or the … view at source ↗
Figure 3
Figure 3. A decomposition unfolds as a sequence of part–remainder transitions. Starting from the TRELLIS. 2 whole-object state (red), each split emits one stored part (green) and carries only the updated remainder forward. In this example, the leaf, arms, body, and legs are removed successively; the rollout terminates when the remainder becomes empty. complete mesh and all of its parts are jointly normalized to the shared obj… view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Qualitative Comparison. All methods in each row use the same conditioning image. The columns show results from Part123, PartField, HoloPart, OmniPart, and SCULPT for four dataset images, one text-to-image example, and one real photograph (bottom row). An asterisk marks…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

56 extracted references · 35 canonical work pages

  1. [1]

    Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

    Structured 3d latents for scalable and versatile 3d generation , author=. Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages=

  2. [2]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , month=

    Native and Compact Structured Latents for 3D Generation , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , month=. 2026 , pages=

    work page 2026

  3. [3]

    arXiv preprint arXiv:2506.16504 , year=

    Hunyuan3D 2.5: Towards High-Fidelity 3D Assets Generation with Ultimate Details , author=. arXiv preprint arXiv:2506.16504 , year=

    arXiv

  4. [4]

    The Eleventh International Conference on Learning Representations , year=

    DreamFusion: Text-to-3D using 2D Diffusion , author=. The Eleventh International Conference on Learning Representations , year=

  5. [5]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , month =

    Lin, Chen-Hsuan and Gao, Jun and Tang, Luming and Takikawa, Towaki and Zeng, Xiaohui and Huang, Xun and Kreis, Karsten and Fidler, Sanja and Liu, Ming-Yu and Lin, Tsung-Yi , title =. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , month =. 2023 , pages =

    work page 2023

  6. [6]

    2022 , eprint=

    Point-E: A System for Generating 3D Point Clouds from Complex Prompts , author=. 2022 , eprint=

    work page 2022

  7. [7]

    2023 , eprint=

    Shap-E: Generating Conditional 3D Implicit Functions , author=. 2023 , eprint=

    work page 2023

  8. [8]

    2024 , url=

    Yichun Shi and Peng Wang and Jianglong Ye and Long Mai and Kejie Li and Xiao Yang , booktitle=. 2024 , url=

    work page 2024

Show all 56 references
  1. [9]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , month =

    Long, Xiaoxiao and Guo, Yuan-Chen and Lin, Cheng and Liu, Yuan and Dou, Zhiyang and Liu, Lingjie and Ma, Yuexin and Zhang, Song-Hai and Habermann, Marc and Theobalt, Christian and Wang, Wenping , title =. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Re...

    2024

  2. [10]

    2024 , url=

    Yicong Hong and Kai Zhang and Jiuxiang Gu and Sai Bi and Yang Zhou and Difan Liu and Feng Liu and Kalyan Sunkavalli and Trung Bui and Hao Tan , booktitle=. 2024 , url=

    2024

  3. [11]

    2024 , isbn =

    Tang, Jiaxiang and Chen, Zhaoxi and Chen, Xiaokang and Wang, Tengfei and Zeng, Gang and Liu, Ziwei , title =. 2024 , isbn =. doi:10.1007/978-3-031-73235-5_1 , booktitle =

    2024 doi

  4. [12]

    arXiv preprint arXiv:2404.07191 , year=

    InstantMesh: Efficient 3D Mesh Generation from a Single Image with Sparse-view Large Reconstruction Models , author=. arXiv preprint arXiv:2404.07191 , year=

    arXiv

  5. [13]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , month=

    Text-Image Conditioned 3D Generation , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , month=. 2026 , pages=

    2026

  6. [14]

    and Yi, Li and Tripathi, Subarna and Guibas, Leonidas J

    Mo, Kaichun and Zhu, Shilin and Chang, Angel X. and Yi, Li and Tripathi, Subarna and Guibas, Leonidas J. and Su, Hao , title =. The IEEE Conference on Computer Vision and Pattern Recognition (CVPR) , month =

  7. [15]

    2024 , eprint=

    SAMPart3D: Segment Any Part in 3D Objects , author=. 2024 , eprint=

    2024

  8. [16]

    Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV) , month =

    Liu, Minghua and Uy, Mikaela Angelina and Xiang, Donglai and Su, Hao and Fidler, Sanja and Sharp, Nicholas and Gao, Jun , title =. Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV) , month =. 2025 , pages =

    2025

  9. [17]

    2025 , eprint=

    P3-SAM: Native 3D Part Segmentation , author=. 2025 , eprint=

    2025

  10. [18]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , month =

    Chen, Minghao and Shapovalov, Roman and Laina, Iro and Monnier, Tom and Wang, Jianyuan and Novotny, David and Vedaldi, Andrea , title =. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , month =. 2025 , pages =

    2025

  11. [19]

    The Fourteenth International Conference on Learning Representations , year=

    HoloPart: Generative 3D Part Amodal Segmentation , author=. The Fourteenth International Conference on Learning Representations , year=

  12. [20]

    The Thirty-ninth Annual Conference on Neural Information Processing Systems , year=

    Efficient Part-level 3D Object Generation via Dual Volume Packing , author=. The Thirty-ninth Annual Conference on Neural Information Processing Systems , year=

  13. [21]

    The Thirty-ninth Annual Conference on Neural Information Processing Systems , year=

    PartCrafter: Structured 3D Mesh Generation via Compositional Latent Diffusion Transformers , author=. The Thirty-ninth Annual Conference on Neural Information Processing Systems , year=

  14. [22]

    2025 , isbn =

    Yang, Yunhan and Zhou, Yufan and Guo, Yuan-Chen and Zou, Zi-Xin and Huang, Yukun and Liu, Ying-Tian and Xu, Hao and Liang, Ding and Cao, Yan-Pei and Liu, Xihui , title =. 2025 , isbn =. doi:10.1145/3757377.3763872 , booktitle =

    2025

  15. [23]

    2024 , isbn =

    Liu, Anran and Lin, Cheng and Liu, Yuan and Long, Xiaoxiao and Dou, Zhiyang and Guo, Hao-Xiang and Luo, Ping and Wang, Wenping , title =. 2024 , isbn =. doi:10.1145/3641519.3657482 , booktitle =

    2024

  16. [24]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , month=

    X-Part: High Fidelity And Structure Coherent Shape Decomposition And Completion , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , month=. 2026 , pages=

    2026

  17. [25]

    Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV) , month =

    Zhang, Lvmin and Rao, Anyi and Agrawala, Maneesh , title =. Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV) , month =. 2023 , pages =

    2023

  18. [26]

    2024 , isbn =

    Mou, Chong and Wang, Xintao and Xie, Liangbin and Wu, Yanze and Zhang, Jian and Qi, Zhongang and Shan, Ying , title =. 2024 , isbn =. doi:10.1609/aaai.v38i5.28226 , booktitle =

    2024 doi

  19. [27]

    2023 , eprint=

    IP-Adapter: Text Compatible Image Prompt Adapter for Text-to-Image Diffusion Models , author=. 2023 , eprint=

    2023

  20. [28]

    The Twelfth International Conference on Learning Representations , year=

    JointNet: Extending Text-to-Image Diffusion for Dense Distribution Modeling , author=. The Twelfth International Conference on Learning Representations , year=

  21. [29]

    Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV) , month =

    Kwon, Byung-Ki and Dai, Qi and Lee, Hyoseok and Luo, Chong and Oh, Tae-Hyun , title =. Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV) , month =. 2025 , pages =

    2025

  22. [30]

    2025 , eprint=

    UniLat3D: Geometry-Appearance Unified Latents for Single-Stage 3D Generation , author=. 2025 , eprint=

    2025

  23. [31]

    Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV) , month =

    Dong, Shaocong and Ding, Lihe and Chen, Xiao and Li, Yaokun and Wang, Yuxin and Wang, Yucheng and Wang, Qi and Kim, Jaehyeok and Gao, Chenjian and Huang, Zhanpeng and Wang, Zibin and Xue, Tianfan and Xu, Dan , title =. Proceedings of the IEEE/CVF International Conference on Co...

    2025

  24. [32]

    The Fourteenth International Conference on Learning Representations , year=

    FullPart: Generating each 3D Part at Full Resolution , author=. The Fourteenth International Conference on Learning Representations , year=

  25. [33]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    Xue, Le and Gao, Mingfei and Xing, Chen and Mart. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  26. [34]

    2023 , eprint=

    SAM3D: Segment Anything in 3D Scenes , author=. 2023 , eprint=

    2023

  27. [35]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , month =

    Yi, Taoran and Fang, Jiemin and Wang, Junjie and Wu, Guanjun and Xie, Lingxi and Zhang, Xiaopeng and Liu, Wenyu and Tian, Qi and Wang, Xinggang , title =. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , month =. 2024 , pages =

    2024

  28. [36]

    2024 , issue_date =

    Yang, Chen and Li, Sikuang and Fang, Jiemin and Liang, Ruofan and Xie, Lingxi and Zhang, Xiaopeng and Shen, Wei and Tian, Qi , title =. 2024 , issue_date =. doi:10.1145/3687759 , journal =

    2024 doi

  29. [37]

    Denoising Diffusion Probabilistic Models , url =

    Ho, Jonathan and Jain, Ajay and Abbeel, Pieter , booktitle =. Denoising Diffusion Probabilistic Models , url =

  30. [38]

    International Conference on Learning Representations , year=

    Denoising Diffusion Implicit Models , author=. International Conference on Learning Representations , year=

  31. [39]

    , title =

    Qi, Charles Ruizhongtai and Yi, Li and Su, Hao and Guibas, Leonidas J. , title =. Proceedings of the 31st International Conference on Neural Information Processing Systems , pages =. 2017 , isbn =

    2017

  32. [40]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , month =

    Liu, Minghua and Zhu, Yinhao and Cai, Hong and Han, Shizhong and Ling, Zhan and Porikli, Fatih and Su, Hao , title =. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , month =. 2023 , pages =

    2023

  33. [41]

    Yuchen Zhou and Jiayuan Gu and Tung Yen Chiang and Fanbo Xiang and Hao Su , booktitle=. Point-. 2025 , url=

    2025

  34. [42]

    Zhe Zhu and Le Wan and Rui Xu and Yiheng Zhang and Honghua Chen and Zhiyang Dou and Cheng Lin and Yuan Liu and Mingqiang Wei , booktitle=. Part. 2026 , url=

    2026

  35. [43]

    Transactions on Machine Learning Research , issn=

    Maxime Oquab and Timoth. Transactions on Machine Learning Research , issn=. 2024 , url=

    2024

  36. [44]

    2025 , eprint=

    DINOv3 , author=. 2025 , eprint=

    2025

  37. [45]

    and Lo, Wan-Yen and Dollar, Piotr and Girshick, Ross , title =

    Kirillov, Alexander and Mintun, Eric and Ravi, Nikhila and Mao, Hanzi and Rolland, Chloe and Gustafson, Laura and Xiao, Tete and Whitehead, Spencer and Berg, Alexander C. and Lo, Wan-Yen and Dollar, Piotr and Girshick, Ross , title =. Proceedings of the IEEE/CVF International ...

    2023

  38. [46]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , month =

    Wu, Xiaoyang and Jiang, Li and Wang, Peng-Shuai and Liu, Zhijian and Liu, Xihui and Qiao, Yu and Ouyang, Wanli and He, Tong and Zhao, Hengshuang , title =. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , month =. 2024 , pages =

    2024

  39. [47]

    Advances in Neural Information Processing Systems , volume=

    Objaverse-XL: A Universe of 10M+ 3D Objects , author=. Advances in Neural Information Processing Systems , volume=

  40. [48]

    2025 , publisher =

    Zhang, Longwen and Zhang, Qixuan and Jiang, Haoran and Bai, Yinuo and Yang, Wei and Xu, Lan and Yu, Jingyi , title =. 2025 , publisher =. doi:10.1145/3730840 , journal =

    2025 doi

  41. [49]

    2026 , pages=

    He, Xufan and Wu, Yushuang and Guo, Xiaoyang and Ye, Chongjie and Zhou, Jiaqing and Hu, Tianlei and Han, Xiaoguang and Du, Dong , booktitle=. 2026 , pages=

    2026

  42. [50]

    Computer Vision -- ECCV 2024 , publisher =

    Ye, Mingqiao and Danelljan, Martin and Yu, Fisher and Ke, Lei , title =. Computer Vision -- ECCV 2024 , publisher =. 2024 , doi =

    2024

  43. [51]

    2026 , url=

    Chunshi Wang and Junliang Ye and Yunhan Yang and Yang Li and Zizhuo Lin and Jun Zhu and Zhuo Chen and Yawei Luo and Chunchao Guo , booktitle=. 2026 , url=

    2026

  44. [52]

    Xiuyu Yang and Yunze Man and Jun-Kun Chen and Yu-Xiong Wang , booktitle=

  45. [53]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , month =

    Yang, Yue and Sun, Fan-Yun and Weihs, Luca and VanderBilt, Eli and Herrasti, Alvaro and Han, Winson and Wu, Jiajun and Haber, Nick and Krishna, Ranjay and Liu, Lingjie and Callison-Burch, Chris and Yatskar, Mark and Kembhavi, Aniruddha and Clark, Christopher , title =. Proceed...

    2024

  46. [54]

    Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV) , month =

    Yan, Siming and Yang, Zhenpei and Ma, Chongyang and Huang, Haibin and Vouga, Etienne and Huang, Qixing , title =. Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV) , month =. 2021 , pages =

    2021

  47. [55]

    and Guibas, Leonidas J

    Mo, Kaichun and Guerrero, Paul and Yi, Li and Su, Hao and Wonka, Peter and Mitra, Niloy J. and Guibas, Leonidas J. , title =. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , month =. 2020 , pages =

    2020

  48. [56]

    2025 , url=

    Minghao Chen and Jianyuan Wang and Roman Shapovalov and Tom Monnier and Hyunyoung Jung and Dilin Wang and Rakesh Ranjan and Iro Laina and Andrea Vedaldi , booktitle=. 2025 , url=

    2025

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.