Pith. sign in

REVIEW 3 major objections 6 minor 65 references

From Air to Wear: Personalized 3D Digital Fashion with AR/VR Immersive 3D Sketching

T0 review · 3 major / 6 minor · reviewed 2026-08-15 · deepseek-v4-flash

Pith's one-line read Drawing a garment in the air with a VR controller can yield a realistic 3D model, according to this paper.

desk verdict The dataset and the input modality are the real contributions; the abstract's 'significantly outperforms' claim is not supported by the paper's own Table I. read the letter →

arxiv 2505.09998 v1 pith:2CV4WFIC submitted 2025-05-15 cs.CV

classification cs.CV
keywords 3DgarmentgenerationVRsketchingsketch-to-3DdiffusionmodelscurriculumlearningpointcloudlatentspaceKO3DClothesdatasetdigitalfashion
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper tries to establish that ordinary people without design training can create personalized 3D clothing by sketching in AR/VR, with an AI model filling in the realism. It argues that imprecise freehand strokes should be treated as a loose condition for a generative model rather than a precise input to regress from, and that a three-stage training scheme—point-cloud diffusion prior, sketch encoder into the same latent space, and joint fine-tuning with adaptive curriculum learning—lets a small paired dataset support this. If true, it matters because it would lower the main barrier to user-generated virtual fashion: instead of modeling software, users only need to draw roughly in the air. The paper's own numbers support the usability claim most strongly: users rated its outputs 4.59/5 for fidelity and 4.62/5 for quality, well above the prior VR-sketch method's 3.28 and 3.39.

What carries the argument

The load-bearing object is a shared 3D latent space: a 1024-dimensional feature embedding produced by a pre-trained point-cloud encoder for garment point clouds and by a Transformer-based sketch encoder for VR sketches, with the conditional diffusion model accepting these embeddings through multi-head cross-attention. The second mechanism is adaptive curriculum learning: a difficulty score per sample based on signed-distance-field prediction errors, a pacing function that feeds easy samples first, and an update rule that re-scores the sample pool during training. Together, these mechanisms let a small paired dataset of 969 sketch–garment pairs drive a high-resolution garment generator.

What would settle it

Re-run the evaluation with a leave-participants-out split of KO3DClothes—no sketcher and no garment shape shared between training and test—and check whether the Chamfer-distance margin over the strongest prior VR-sketch baseline persists.

Watch

Extended reading notes

Core claim

On its own terms, the paper's central claim is that user-created 3D VR sketches—imprecise, incomplete, and stylistically varied—carry enough information to condition a generative model into producing realistic 3D garments. The system, which the paper calls Deep3DVRSketch+, achieves this by pre-training a coarse-to-fine conditional diffusion model on garment point clouds, freezing it while training a Transformer-based sketch encoder into the same latent space, then jointly fine-tuning both stages with an adaptive curriculum that orders samples from easy to hard. On the introduced KO3DClothes dataset of 969 paired sketches and garments, the method reports a Chamfer distance of 0.0597, marginally better than the strongest prior VR-sketch baseline (0.0606) and clearly better than a normalizing-flow baseline (0.0820), while IoU is comparable (0.3190 vs 0.3252 and 0.3188). In a 15-designer user study, fidelity and quality mean opinion scores are well above the prior VR-sketch method, and controllability and usefulness beat a 2D-sketch baseline by large margins. The paper's framing is that the system targets overall garment shape—sleeves, length, silhouette—while leaving wrinkles and dynamic folds to downstream clothing simulation.

Load-bearing premise

The comparison assumes the 8:2 training/test split prevents the ten participants' individual drawing styles from appearing in both halves; if test sketches come from the same people who produced training sketches, the reported gains may reflect memorized stroke habits rather than general understanding of freehand input.

Editorial extensions

If this is right

  • A non-designer wearing a VR headset can produce a plausible personalized 3D garment from rough freehand strokes, removing the main technical skill barrier to virtual fashion creation.
  • A small paired dataset of 969 sketch–garment samples is sufficient to train a usable sketch-conditioned garment generator when a point-cloud diffusion prior and curriculum learning are used.
  • The system targets overall garment shape—silhouette, sleeve presence, and sleeve length—rather than fine wrinkles, on the expectation that downstream cloth simulation adds realistic dynamic folds.
  • Compared against two prior sketch-to-shape baselines, the method reports the lowest Chamfer distance and the best user ratings for fidelity and quality, supporting the viability of VR sketches as a design input.
  • The new KO3DClothes dataset of paired user-created VR sketches and 3D garments provides a benchmark resource for future sketch-based garment generation work.

Reading between the lines

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

  • A direct extension would be leave-one-participant-out evaluation: the paper's 8:2 split is not stated to separate sketching styles, so holding out entire users would test whether the model truly generalizes to unfamiliar stroke habits.
  • Because the sketch is treated as a loose condition rather than a precise contour, the same two-stage latent-alignment recipe could transfer to other shape domains by swapping the pre-trained point-cloud prior, such as furniture or avatars.
  • The reported margin over the strongest prior baseline is small on Chamfer distance; an implication is that the usability advantage, not geometric accuracy, may be the more robust selling point worth reproducing.
  • A testable extension is to measure per-sketch diversity: if the loose condition allows several plausible garments per sketch, the method could support interactive design exploration rather than single-output reconstruction.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 6 minor

Summary. The paper presents a sketch-driven 3D garment generation framework for AR/VR users. The authors introduce KO3DClothes, a dataset of 969 paired 3D garment meshes and human-drawn 3D VR sketches collected from 10 non-professional participants, and a three-stage training strategy: pre-training a conditional diffusion model on point-cloud shapes, training a Transformer sketch encoder in a shared latent space, and jointly fine-tuning both with an adaptive curriculum learning schedule. The method is evaluated against 3DSketch2Shape and Deep3DVRSketch using voxel IoU and Chamfer Distance, together with two user studies. The central claim is that the system interprets imprecise free-hand input, produces realistic personalized garments, and significantly outperforms existing baselines in both fidelity and usability.

Significance. If the claims were fully supported, the paper would make a useful contribution: it addresses a practical data-scarcity problem in sketch-based garment generation, contributes a new paired dataset of real user VR sketches and garment meshes, and combines a point-cloud diffusion prior, shared-latent alignment, and curriculum learning in a way that could be reused by other sketch-to-shape tasks. The ablation study is a strength because it isolates the role of the point-cloud prior and curriculum learning, and the comparison against 3DSketch2Shape and Deep3DVRSketch is methodologically sensible. However, the headline quantitative claim is not supported by the numbers in Table I: on IoU the proposed method is worse than Deep3DVRSketch, and on Chamfer Distance the improvement is only 0.0009 with no error bars or significance tests. The data split is not documented as participant-disjoint, which threatens the generalization claim, and the user studies are not statistically grounded. These are load-bearing issues for the paper's main conclusion, but they appear addressable with additional analysis, statistical reporting, and a more carefully controlled evaluation protocol.

major comments (3)
  1. [V.B, Table I] The abstract claims that the method 'significantly outperforms existing baselines in both fidelity and usability,' but Table I shows IoU of 0.3190 for the proposed method versus 0.3252 for Deep3DVRSketch, and a Chamfer Distance improvement of only 0.0009 (0.0597 versus 0.0606). No standard deviations, confidence intervals, significance tests, or per-sample paired comparisons are reported. On the reported evidence, the method is not better than Deep3DVRSketch on IoU, and the CD margin is within the range that could arise from run-to-run noise. This unsupported 'significant' claim needs to be either substantiated with statistical evidence or removed.
  2. [V.A and III] The train/test split is described only as an 8:2 ratio of the KO3DClothes dataset, while Section III states that all sketches were drawn by only 10 participants. The paper does not state that test sketches were drawn by held-out participants, nor that the split is garment-disjoint or style-stratified. If the same participants' drawing styles appear in both training and test sets, the reported metrics, including the user-study preference, may reflect memorization of a small number of personal stroke habits rather than generalization to arbitrary users. The authors should report a participant-disjoint evaluation or otherwise demonstrate that drawing style does not leak across the split.
  3. [V.C, Tables II and III] The user study is described as 15 designers rating 12 generated results, but the paper does not report how many tasks were rated per method, whether the raters were the same 10 participants who created the training sketches, whether the raters were blinded to the method, or any inter-rater agreement. No significance or effect-size statistics are given for the MOS differences. The reported MOS gap (4.59 versus 3.28 for fidelity) could be affected by rater bias, task difficulty imbalance, or style leakage from the training set. A controlled protocol with independent raters, balanced tasks, and paired statistical tests is required to support the usability claim.
minor comments (6)
  1. [Abstract and I] The phrase 'significantly outperforms' appears in the abstract and introduction, but the quantitative results in Table I do not support a significance claim; please rephrase these statements to match the evidence actually presented.
  2. [IV.B] The section heading 'The Proposed Multi-stage Training Steategy' contains a typo; it should read 'Strategy.'
  3. [I] The sentence 'Just as beginners in sketching typically start to learn with with simple and flexible shapes' contains a duplicated 'with.'
  4. [Fig. 1] The figure label 'DeepVRSketch+' is inconsistent with the method name 'Deep3DVRSketch+' used in the rest of the paper; please unify the terminology.
  5. [V.A] The text refers to a 'KO3DClothes+ dataset' but the dataset is introduced as 'KO3DClothes' in Section III; please use one name consistently.
  6. [IV.C] In Equation (2), the definition of y as 'the SDF value corresponding to the hand-drawn sketch' is unclear because the input is a point cloud; the authors should explain how SDF values are assigned to sketch points and how the difficulty score is computed from them.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the reported gains are empirical comparisons to baselines, not consequences of the method's own definitions or fitted parameters.

full rationale

The paper's central claim is an empirical comparison: Table I reports Chamfer Distance and IoU for 'Ours' versus 3DSketch2Shape and Deep3DVRSketch, with all baselines retrained on the same KO3DClothes dataset. The method's components (conditional diffusion, a sketch encoder mapped into a shared latent space, and curriculum learning with formulas (2)-(4)) are not used to define the evaluation metrics, and no equation in the paper reduces the reported test-set numbers to the training loss or to a hand-set hyperparameter. The self-citations, including Deep3DVRSketch [50] as a baseline, the data-collection protocol adapted from [50],[51], and the MOS-evaluation reference to prior work by the same group, are used as benchmarks or procedural precedents rather than as load-bearing justification for the central result. The potential leakage from an 8:2 split that is not stated to be participant-disjoint, and the small CD margin with worse IoU, are empirical validity concerns, not circular reasoning. The derivation is self-contained against external evaluation, so no circular step is present.

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

Six hyperparameters of the curriculum (alpha, beta, inv, p0, q, r0) are hand-set; three are reported (p0=0.2, q=1.9, r0=1), three are not. The central claim relies on the domain assumption that DeepFashion3D point clouds are a sufficient shape prior and that 969 sketches from 10 amateurs represent ordinary-user input. No new physical or conceptual entities are postulated; the KO3DClothes dataset is a collected resource, not an invented construct.

free parameters (6)
  • alpha (difficulty coefficient) = not reported
    Controls how strongly sign mismatch weights difficulty in the curriculum scoring formula, Eq. (2). Hand-set and never varied or reported.
  • beta (difficulty update speed) = not reported
    Controls the speed of difficulty score updates in Eq. (3). Hand-set and never reported.
  • inv (difficulty update frequency) = not reported
    Controls how often difficulty scores are updated in Eq. (3). Hand-set and never reported.
  • p0 (initial pacing ratio) = 0.2
    Initial sample ratio in the pacing function, Eq. (4). Hand-set in Section IV.C.3.
  • q (pacing growth rate) = 1.9
    Controls the speed of sample ratio growth in Eq. (4). Hand-set in Section IV.C.3.
  • r0 (pacing growth step) = 1
    Controls the frequency of sample ratio growth in Eq. (4). Hand-set in Section IV.C.3.
assumptions (4)
  • domain assumption A pretrained conditional diffusion model on DeepFashion3D point clouds provides a sufficient shape prior for garment generation from VR sketches.
    Stage 1 (Section IV.B.1) builds the generator from this prior; the ablation shows performance drops substantially without it, so the central result depends on it.
  • domain assumption Hand-drawn VR sketches by 10 non-professionals, sampled to 4096 points and manually aligned, capture the distribution of ordinary-user input.
    Section III makes dataset construction choices on this basis; the evaluation claims rest on this representativeness.
  • ad hoc to paper The latent spaces of Uni3D point-cloud features and the Transformer sketch encoder are alignable with 969 paired samples.
    Section IV.B.2 and B.3 posit that sketch features can be mapped into the pre-trained point-cloud latent space; if alignment is not achievable at this data scale the method fails.
  • standard math DDPM denoising objective (Eq. 1) trains a valid generative model.
    Standard diffusion training objective from [53]; accepted background for the method.

how reviews work

0 comments
Cite this review

Pith. "Pith review of From Air to Wear: Personalized 3D Digital Fashion with AR/VR Immersive 3D Sketching." pith.science (2026). https://pith.science/paper/2CV4WFIC

@misc{pith2026250509998,
  author       = {Pith},
  title        = {Pith review of: From Air to Wear: Personalized 3D Digital Fashion with AR/VR Immersive 3D Sketching},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/2CV4WFIC}},
  note         = {Machine review of arXiv:2505.09998}
}
read the original abstract

In the era of immersive consumer electronics, such as AR/VR headsets and smart devices, people increasingly seek ways to express their identity through virtual fashion. However, existing 3D garment design tools remain inaccessible to everyday users due to steep technical barriers and limited data. In this work, we introduce a 3D sketch-driven 3D garment generation framework that empowers ordinary users - even those without design experience - to create high-quality digital clothing through simple 3D sketches in AR/VR environments. By combining a conditional diffusion model, a sketch encoder trained in a shared latent space, and an adaptive curriculum learning strategy, our system interprets imprecise, free-hand input and produces realistic, personalized garments. To address the scarcity of training data, we also introduce KO3DClothes, a new dataset of paired 3D garments and user-created sketches. Extensive experiments and user studies confirm that our method significantly outperforms existing baselines in both fidelity and usability, demonstrating its promise for democratized fashion design on next-generation consumer platforms.

Figures

Figures reproduced from arXiv: 2505.09998 by the authors.

Figure 1
Figure 1. In this work, we propose a novel method that allows everyday users to create personalized 3D garments by simply sketching in immersive [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. The visualization of hand-drawn 3D sketch samples from [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. The Overview of Deep3DVRSketch+. (a) Pre-training a conditional diffusion model by sampling ground truth (GT) point clouds. (b) Fine-tuning the [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Comparison with the existing state-of-the-art methods. [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 5
Figure 5. Figure 5: Qualitative Evaluation for Ablation Studies. [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

65 extracted references · 45 canonical work pages

  1. [11]

    Pamir: Parametric model- conditioned implicit representation for image-based human reconstruc- tion,

    Z. Zheng, T. Yu, Y . Liu, and Q. Dai, “Pamir: Parametric model- conditioned implicit representation for image-based human reconstruc- tion,” IEEE transactions on pattern analysis and machine intelligence , vol. 44, no. 6, pp. 3170–3184, 2021

  2. [1]

    Mannequin2real: a two-stage generation framework for transforming mannequin images into photorealistic model images for clothing display,

    H. Zhang, X. Mu, G. Li, Z. Xu, X. Yu, and J. Ma, “Mannequin2real: a two-stage generation framework for transforming mannequin images into photorealistic model images for clothing display,” IEEE Transac- tions on Consumer Electronics , 2024

  3. [2]

    Towards high-fidelity 3d virtual try-on via global collaborative modeling,

    X. Hu, C. Fang, K. Yang, J. Liang, R. Luo, and T. Peng, “Towards high-fidelity 3d virtual try-on via global collaborative modeling,” IEEE Transactions on Consumer Electronics , 2024

  4. [3]

    The democratization of luxury–a new form of luxury,

    K. Pla ˙zyk, “The democratization of luxury–a new form of luxury,” Studia University of Economics in Katowice , pp. 158–165, 2015

  5. [4]

    Hopkins, Fashion design: The complete guide

    J. Hopkins, Fashion design: The complete guide . Bloomsbury Publish- ing, 2021

  6. [5]

    Experimental evaluation of sketching on surfaces in vr

    R. Arora, R. H. Kazi, F. Anderson, T. Grossman, K. Singh, and G. W. Fitzmaurice, “Experimental evaluation of sketching on surfaces in vr.” in CHI 2017, vol. 17, 2017, pp. 5643–5654

  7. [6]

    Sketchsampler: Sketch-based 3d reconstruction via view-dependent depth sampling,

    C. Gao, Q. Yu, L. Sheng, Y .-Z. Song, and D. Xu, “Sketchsampler: Sketch-based 3d reconstruction via view-dependent depth sampling,” in European Conference on Computer Vision . Springer, 2022, pp. 464– 479

  8. [7]

    Holosketch: a virtual reality sketching/animation tool,

    M. F. Deering, “Holosketch: a virtual reality sketching/animation tool,” ACM Transactions on Computer-Human Interaction (TOCHI) , vol. 2, no. 3, pp. 220–238, 1995

Show all 65 references
  1. [8]

    Cavepainting: A fully immersive 3d artistic medium and interactive experience,

    D. F. Keefe, D. A. Feliz, T. Moscovich, D. H. Laidlaw, and J. J. LaViola Jr, “Cavepainting: A fully immersive 3d artistic medium and interactive experience,” in Proceedings of the 2001 symposium on Interactive 3D graphics , 2001, pp. 85–93

  2. [9]

    Mobi3dsketch: 3d sketching in mobile ar,

    K. C. Kwan and H. Fu, “Mobi3dsketch: 3d sketching in mobile ar,” in CHI 2019, 2019, pp. 1–11

  3. [10]

    Model- guided 3d sketching,

    P. Xu, H. Fu, Y . Zheng, K. Singh, H. Huang, and C.-L. Tai, “Model- guided 3d sketching,” IEEE Transactions on Visualization and Computer Graphics, vol. 25, no. 10, pp. 2927–2939, 2018

  4. [12]

    Sketch2cloth: Sketch-based 3d garment generation with unsigned distance fields,

    Y . He, H. Xie, and K. Miyata, “Sketch2cloth: Sketch-based 3d garment generation with unsigned distance fields,” in 2023 Nicograph Interna- tional (NicoInt). IEEE, 2023, pp. 38–45

  5. [13]

    Weakly supervised joint transfer and regression of textures for 3d human reconstruction,

    F. Zhao, S. Liao, J. Huo, Z. Huo, W. Wang, J. Han, and C. Shan, “Weakly supervised joint transfer and regression of textures for 3d human reconstruction,” IEEE Transactions on Consumer Electronics , 2024

  6. [14]

    Bcnet: Learning body and cloth shape from a single image,

    B. Jiang, J. Zhang, Y . Hong, J. Luo, L. Liu, and H. Bao, “Bcnet: Learning body and cloth shape from a single image,” in Computer Vision–ECCV 2020: 16th European Conference, Glasgow, UK, August 23–28, 2020, Proceedings, Part XX 16 . Springer, 2020, pp. 18–35

  7. [15]

    Registering explicit to implicit: Towards high-fidelity garment mesh reconstruction from single images,

    H. Zhu, L. Qiu, Y . Qiu, and X. Han, “Registering explicit to implicit: Towards high-fidelity garment mesh reconstruction from single images,” https://arxiv.org/abs/2203.15007, 2022. [Online]. Available: https://arxiv.org/abs/2203.15007

  8. [16]

    xcloth: Extracting template-free textured 3d clothes from a monocular image,

    A. Srivastava, C. Pokhariya, S. S. Jinka, and A. Sharma, “xcloth: Extracting template-free textured 3d clothes from a monocular image,” in Proceedings of the 30th ACM International Conference on Multimedia , 2022, pp. 2504–2512

  9. [17]

    Multi- garment net: Learning to dress 3d people from images,

    B. L. Bhatnagar, G. Tiwari, C. Theobalt, and G. Pons-Moll, “Multi- garment net: Learning to dress 3d people from images,” in Proceedings of the IEEE/CVF international conference on computer vision , 2019, pp. 5420–5430

  10. [19]

    Pifu: Pixel-aligned implicit function for high-resolution clothed human digitization,

    S. Saito, Z. Huang, R. Natsume, S. Morishima, A. Kanazawa, and H. Li, “Pifu: Pixel-aligned implicit function for high-resolution clothed human digitization,” in Proceedings of the IEEE/CVF international conference on computer vision , 2019, pp. 2304–2314

  11. [20]

    Pifuhd: Multi-level pixel- aligned implicit function for high-resolution 3d human digitization,

    S. Saito, T. Simon, J. Saragih, and H. Joo, “Pifuhd: Multi-level pixel- aligned implicit function for high-resolution 3d human digitization,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2020, pp. 84–93

  12. [21]

    Econ: Explicit clothed humans optimized via normal integration,

    Y . Xiu, J. Yang, X. Cao, D. Tzionas, and M. J. Black, “Econ: Explicit clothed humans optimized via normal integration,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , 2023, pp. 512–523

  13. [22]

    Computational pattern making from 3d garment models

    N. Pietroni, C. Dumery, R. Falque, M. Liu, T. A. Vidal-Calleja, and O. Sorkine-Hornung, “Computational pattern making from 3d garment models.” ACM Trans. Graph., vol. 41, no. 4, pp. 157–1, 2022

  14. [23]

    Generating datasets of 3d garments with sewing patterns,

    M. Korosteleva and S.-H. Lee, “Generating datasets of 3d garments with sewing patterns,” arXiv preprint arXiv:2109.05633 , 2021

  15. [24]

    Smplicit: Topology-aware generative model for clothed peo- JOURNAL OF LATEX CLASS FILES, VOL. 14, NO. 8, AUGUST 2021 8 ple,

    E. Corona, A. Pumarola, G. Alenya, G. Pons-Moll, and F. Moreno- Noguer, “Smplicit: Topology-aware generative model for clothed peo- JOURNAL OF LATEX CLASS FILES, VOL. 14, NO. 8, AUGUST 2021 8 ple,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recogn...

  16. [25]

    Deepcloth: Neural garment representation for shape and style editing,

    Z. Su, T. Yu, Y . Wang, and Y . Liu, “Deepcloth: Neural garment representation for shape and style editing,”IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 45, no. 2, pp. 1581–1593, 2022

  17. [26]

    Surf-d: High-quality surface gen- eration for arbitrary topologies using diffusion models,

    Z. Yu, Z. Dou, X. Long, C. Lin, Z. Li, Y . Liu, N. M ¨uller, T. Komura, M. Habermann, C. Theobalt et al. , “Surf-d: High-quality surface gen- eration for arbitrary topologies using diffusion models,” arXiv preprint arXiv:2311.17050, 2023

  18. [27]

    Garment ideation: Iterative view-aware sketch-based garment model- ing,

    P. N. Chowdhury, T. Wang, D. Ceylan, Y .-Z. Song, and Y . Gryaditskaya, “Garment ideation: Iterative view-aware sketch-based garment model- ing,” in 2022 International Conference on 3D Vision (3DV) . IEEE, 2022, pp. 22–31

  19. [28]

    Doodle your 3d: From abstract free- hand sketches to precise 3d shapes,

    H. Bandyopadhyay, S. Koley, A. Das, A. Sain, P. N. Chowdhury, T. Xi- ang, A. K. Bhunia, and Y .-Z. Song, “Doodle your 3d: From abstract free- hand sketches to precise 3d shapes,” arXiv preprint arXiv:2312.04043 , 2023

  20. [29]

    Sdfusion: Multimodal 3d shape completion, reconstruction, and gener- ation,

    Y .-C. Cheng, H.-Y . Lee, S. Tulyakov, A. G. Schwing, and L.-Y . Gui, “Sdfusion: Multimodal 3d shape completion, reconstruction, and gener- ation,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2023, pp. 4456–4465

  21. [30]

    A diffusion-refinement model for sketch-to-point modeling,

    D. Kong, Q. Wang, and Y . Qi, “A diffusion-refinement model for sketch-to-point modeling,” in Proceedings of the Asian Conference on Computer Vision, 2022, pp. 1522–1538

  22. [31]

    3d-ldm: Neural implicit 3d shape generation with latent diffusion models,

    G. Nam, M. Khlifi, A. Rodriguez, A. Tono, L. Zhou, and P. Guerrero, “3d-ldm: Neural implicit 3d shape generation with latent diffusion models,” arXiv preprint arXiv:2212.00842 , 2022

  23. [32]

    Clip-forge: Towards zero-shot text-to-shape generation,

    A. Sanghi, H. Chu, J. G. Lambourne, Y . Wang, C.-Y . Cheng, M. Fumero, and K. R. Malekshan, “Clip-forge: Towards zero-shot text-to-shape generation,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2022, pp. 18 603–18 613

  24. [33]

    Generative ai meets 3d: A survey on text-to-3d in aigc era,

    C. Li, C. Zhang, A. Waghwase, L.-H. Lee, F. Rameau, Y . Yang, S.-H. Bae, and C. S. Hong, “Generative ai meets 3d: A survey on text-to-3d in aigc era,” arXiv preprint arXiv:2305.06131 , 2023

  25. [34]

    Shapecrafter: A recursive text-conditioned 3d shape generation model,

    R. Fu, X. Zhan, Y . Chen, D. Ritchie, and S. Sridhar, “Shapecrafter: A recursive text-conditioned 3d shape generation model,” Advances in Neural Information Processing Systems , vol. 35, pp. 8882–8895, 2022

  26. [35]

    Shapescaffolder: Structure-aware 3d shape generation from text,

    X. Tian, Y .-L. Yang, and Q. Wu, “Shapescaffolder: Structure-aware 3d shape generation from text,” in Proceedings of the IEEE/CVF International Conference on Computer Vision , 2023, pp. 2715–2724

  27. [36]

    One-2-3-45: Any single image to 3d mesh in 45 seconds without per-shape optimization,

    M. Liu, C. Xu, H. Jin, L. Chen, Z. Xu, H. Su et al., “One-2-3-45: Any single image to 3d mesh in 45 seconds without per-shape optimization,” arXiv preprint arXiv:2306.16928 , 2023

  28. [37]

    Mvdream: Multi- view diffusion for 3d generation,

    Y . Shi, P. Wang, J. Ye, M. Long, K. Li, and X. Yang, “Mvdream: Multi- view diffusion for 3d generation,” arXiv preprint arXiv:2308.16512 , 2023

  29. [38]

    Sketch2model: View-aware 3d modeling from single free-hand sketches,

    S.-H. Zhang, Y .-C. Guo, and Q.-W. Gu, “Sketch2model: View-aware 3d modeling from single free-hand sketches,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2021, pp. 6012–6021

  30. [39]

    Sketch2mesh: Re- constructing and editing 3d shapes from sketches,

    B. Guillard, E. Remelli, P. Yvernay, and P. Fua, “Sketch2mesh: Re- constructing and editing 3d shapes from sketches,” in Proceedings of the IEEE/CVF International Conference on Computer Vision , 2021, pp. 13 023–13 032

  31. [40]

    Deep3dsketch: 3d modeling from free-hand sketches with view-and structural-aware adversarial training,

    T. Chen, C. Fu, L. Zhu, P. Mao, J. Zhang, Y . Zang, and L. Sun, “Deep3dsketch: 3d modeling from free-hand sketches with view-and structural-aware adversarial training,” arXiv preprint arXiv:2312.04435, 2023

  32. [41]

    Sketch-based modeling: A survey,

    L. Olsen, F. F. Samavati, M. C. Sousa, and J. A. Jorge, “Sketch-based modeling: A survey,”Computers & Graphics, vol. 33, no. 1, pp. 85–103, 2009

  33. [42]

    Sketch-based interaction and modeling: where do we stand?

    A. Bonnici, A. Akman, G. Calleja, K. P. Camilleri, P. Fehling, A. Fer- reira, F. Hermuth, J. H. Israel, T. Landwehr, J. Liu et al., “Sketch-based interaction and modeling: where do we stand?” AI EDAM, vol. 33, no. 4, pp. 370–388, 2019

  34. [43]

    3d shape reconstruction from sketches via multi-view convolutional networks,

    Z. Lun, M. Gadelha, E. Kalogerakis, S. Maji, and R. Wang, “3d shape reconstruction from sketches via multi-view convolutional networks,” in 2017 International Conference on 3D Vision (3DV) . IEEE, 2017, pp. 67–77

  35. [44]

    Robust flow- guided neural prediction for sketch-based freeform surface modeling,

    C. Li, H. Pan, Y . Liu, X. Tong, A. Sheffer, and W. Wang, “Robust flow- guided neural prediction for sketch-based freeform surface modeling,” ACM Transactions on Graphics (TOG) , vol. 37, no. 6, pp. 1–12, 2018

  36. [45]

    3d shape reconstruction from free-hand sketches,

    J. Wang, J. Lin, Q. Yu, R. Liu, Y . Chen, and S. X. Yu, “3d shape reconstruction from free-hand sketches,” in European Conference on Computer Vision. Springer, 2022, pp. 184–202

  37. [46]

    Deep sketch- based modeling: Tips and tricks,

    Y . Zhong, Y . Gryaditskaya, H. Zhang, and Y .-Z. Song, “Deep sketch- based modeling: Tips and tricks,” in 2020 International Conference on 3D Vision (3DV). IEEE, 2020, pp. 543–552

  38. [47]

    Deep3dsketch+: Rapid 3d modeling from single free-hand sketches,

    T. Chen, C. Fu, Y . Zang, L. Zhu, J. Zhang, P. Mao, and L. Sun, “Deep3dsketch+: Rapid 3d modeling from single free-hand sketches,” in International Conference on Multimedia Modeling . Springer, 2023, pp. 16–28

  39. [48]

    Deep3dsketch+: Obtaining customized 3d model by single free-hand sketch through deep learning,

    Y . Zang, C. Fu, T. Chen, Y . Hu, Q. Liu, and W. Hu, “Deep3dsketch+: Obtaining customized 3d model by single free-hand sketch through deep learning,” arXiv preprint arXiv:2310.18609 , 2023

  40. [49]

    Deep3dsketch+ \+: High-fidelity 3d modeling from single free-hand sketches,

    Y . Zang, C. Ding, T. Chen, P. Mao, and W. Hu, “Deep3dsketch+ \+: High-fidelity 3d modeling from single free-hand sketches,” arXiv preprint arXiv:2310.18178, 2023

  41. [50]

    Rapid 3d model generation with intuitive 3d input,

    T. Chen, C. Ding, S. Zhang, C. Yu, Y . Zang, Z. Li, S. Peng, and L. Sun, “Rapid 3d model generation with intuitive 3d input,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2024, pp. 12 554–12 564

  42. [51]

    Fine- grained vr sketching: Dataset and insights,

    L. Luo, Y . Gryaditskaya, Y . Yang, T. Xiang, and Y .-Z. Song, “Fine- grained vr sketching: Dataset and insights,” in 2021 International Conference on 3D Vision (3DV) . IEEE, 2021, pp. 1003–1013

  43. [52]

    Locally attentional sdf diffusion for controllable 3d shape generation,

    X.-Y . Zheng, H. Pan, P.-S. Wang, X. Tong, Y . Liu, and H.-Y . Shum, “Locally attentional sdf diffusion for controllable 3d shape generation,” arXiv preprint arXiv:2305.04461 , 2023

  44. [53]

    Denoising diffusion probabilistic models,

    J. Ho, A. Jain, and P. Abbeel, “Denoising diffusion probabilistic models,” Advances in neural information processing systems , vol. 33, pp. 6840– 6851, 2020

  45. [54]

    Classifier-free diffusion guidance,

    J. Ho and T. Salimans, “Classifier-free diffusion guidance,” arXiv preprint arXiv:2207.12598, 2022

  46. [55]

    Uni3d: Exploring unified 3d representation at scale,

    J. Zhou, J. Wang, B. Ma, Y .-S. Liu, T. Huang, and X. Wang, “Uni3d: Exploring unified 3d representation at scale,” arXiv preprint arXiv:2310.06773, 2023

  47. [56]

    Pretraining is all you need for image-to-image translation,

    T. Wang, T. Zhang, B. Zhang, H. Ouyang, D. Chen, Q. Chen, and F. Wen, “Pretraining is all you need for image-to-image translation,” arXiv preprint arXiv:2205.12952 , 2022

  48. [57]

    Curriculum deepsdf,

    Y . Duan, H. Zhu, H. Wang, L. Yi, R. Nevatia, and L. J. Guibas, “Curriculum deepsdf,” in Computer Vision–ECCV 2020: 16th European Conference, Glasgow, UK, August 23–28, 2020, Proceedings, Part VIII

  49. [58]

    Adaptive curriculum learning,

    Y . Kong, L. Liu, J. Wang, and D. Tao, “Adaptive curriculum learning,” in Proceedings of the IEEE/CVF International Conference on Computer Vision, 2021, pp. 5067–5076

  50. [59]

    Springer, 2020, pp. 51–67

  51. [60]

    Decoupled weight decay regularization,

    I. Loshchilov and F. Hutter, “Decoupled weight decay regularization,” arXiv preprint arXiv:1711.05101 , 2017

  52. [61]

    Adam: A method for stochastic optimization,

    D. P. Kingma and J. Ba, “Adam: A method for stochastic optimization,” arXiv preprint arXiv:1412.6980 , 2014

  53. [62]

    I lead, you help but only with enough details: Understanding user experience of co-creation with artificial intelligence,

    C. Oh, J. Song, J. Choi, S. Kim, S. Lee, and B. Suh, “I lead, you help but only with enough details: Understanding user experience of co-creation with artificial intelligence,” in CHI 2018, 2018, pp. 1–13

  54. [63]

    3d vr sketch guided 3d shape prototyping and exploration,

    L. Luo, P. N. Chowdhury, T. Xiang, Y .-Z. Song, and Y . Gryaditskaya, “3d vr sketch guided 3d shape prototyping and exploration,” in Proceedings of the IEEE/CVF International Conference on Computer Vision , 2023, pp. 9267–9276

  55. [64]

    Magic3dsketch: Create colorful 3d models from sketch-based 3d modeling guided by text and language-image pre-training,

    Y . Zang, Y . Han, C. Ding, J. Zhang, and T. Chen, “Magic3dsketch: Create colorful 3d models from sketch-based 3d modeling guided by text and language-image pre-training,” arXiv preprint arXiv:2407.19225, 2024

  56. [65]

    Albert and T

    B. Albert and T. Tullis, Measuring the User Experience: Collecting, Analyzing, and Presenting UX Metrics . Morgan Kaufmann, 2022

  57. [2020]

    Available: https://arxiv.org/abs/2003.12753

    [Online]. Available: https://arxiv.org/abs/2003.12753

Pith tools

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