Pith. sign in

REVIEW 4 major objections 6 minor 45 references

VQ-SGen: A Vector Quantized Stroke Representation for Creative Sketch Generation

T0 review · 4 major / 6 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read VQ-SGen claims that creative sketch generation works best stroke-by-stroke, encoding each stroke as a vector-quantized token and generating sketches with a cascaded autoregressive Transformer that outperforms pixel-based methods on…

desk verdict A clean stroke-level VQ representation with a solid method, but the evaluation needs error bars and a broader user study before 'surpasses SOTA' sticks. read the letter →

arxiv 2411.16446 v3 pith:4M2LWGB5 submitted 2024-11-25 cs.CV cs.GR

Jiawei Wang , Zhiming Cui , Changjian Li This is my paper · ORCID
classification cs.CVcs.GR
keywords creativesketchgenerationvectorquantizationstrokerepresentationautoregressiveTransformertext-to-sketchcompletiondatasetdiscretecodebook
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

VQ-SGen sets out to show that creative sketch generation should be done at the level of individual strokes rather than pixels or whole parts. The paper proposes a two-stage pipeline: first, it decouples each stroke's shape from its location and encodes both into separate vector-quantized codebooks; second, a cascaded autoregressive Transformer generates a sketch as a sequence of stroke labels, shape codes, and location codes. On the CreativeSketch benchmark, this outperforms existing methods on FID and generation diversity (FID 15.78 vs 17.48 on Creative Birds, 17.61 vs 20.43 on Creative Creatures) and produces sketches with fewer of the blurry regions and broken strokes seen in earlier work. If true, this means a discrete stroke-token representation is a strong basis for sketch generation, and it also opens label-, text-, and completion-conditioned drawing from the same model.

What carries the argument

The load-bearing object is the vector-quantized stroke representation paired with a cascaded autoregressive Transformer. A stroke is rasterized as a 256x256 image, its axis-aligned bounding box coordinates $(w/2, h/2, x, y)$ define location, and the shape is the stroke translated to image center; two codebooks (each 8192 entries, 512 dimensions in the final configuration) quantize shape and location separately. Two Transformer decoders then predict, in order, the next stroke's label and its shape/location codes, sampling in the discrete code space instead of raw pixels. Training uses reconstruction, codebook, and commitment losses from VQ-VAE, with CoordConv layers and distance-field supervision in the stroke autoencoder. The decoupling forces the shape codebook to specialize on stroke geometry and the position codebook on inter-stroke layout, which the ablations show is what drives the FID improvement (15.78 vs 17.14 without decoupling on Creative Birds).

What would settle it

A concrete check: re-run the CreativeSketch comparison with an evaluator not trained on QuickDraw (for instance a text-aligned visual encoder or a sketch-specific perceptual metric) and see whether VQ-SGen still beats DoodleFormer on FID; alternatively, run a human creativity-preference study on pairs matched for recognizability to see whether the fine-grained stroke quality the paper claims is perceptible to raters.

Watch

Extended reading notes

Core claim

The paper's central claim is that a sketch is best modeled as a sequence of stroke entities, each represented by a compact discrete code for its shape and another for its position, and then regenerated autoregressively. VQ-SGen first translates each rasterized stroke so its bounding box is centered, records the box as the location code, learns a VQ codebook for stroke shapes and another for locations, and then trains two cascaded Transformer decoders: one predicts the next stroke's semantic label, the other predicts its shape and position codes conditioned on that label and on all previous strokes. The authors report that this discrete, decoupled representation outperforms pixel-level and point-level baselines on the CreativeSketch dataset, that the learned shape code space clusters by semantic stroke role (beak, wing, tail, eye, body) without semantic supervision, and that the same machinery supports class-label, text, and completion conditioning. The argument is that the VQ representation compresses away redundancy while preserving stroke shape variations, giving the generator a semantically aware, compact space to sample from.

Load-bearing premise

The load-bearing premise is that the Inception feature space trained on QuickDraw doodles, together with the FID, GD, CS, and SDS metrics computed in it, actually tracks human judgments of creativity and sketch quality; if it does not, the numerical superiority over prior methods would not establish the paper's conclusion.

Editorial extensions

If this is right

  • If the claim holds, stroke-token generation replaces pixel-level and part-level generation as the stronger approach on CreativeSketch, with lower FID and higher diversity.
  • Class-label and text conditioning work by swapping the start token, so one trained generator covers unconditional, conditional, and completion modes without a new architecture.
  • Sketch completion reduces to seeding the first stroke, and the paper's comparison shows this beats the pixel-level baselines at completing a sketch from an initial mark.
  • The near-parity of the w/o T_l ablation indicates the method does not depend on stroke labels, so it transfers to unlabeled datasets like the QuickDraw subset used in the application.
  • The semantic clustering that emerges in the shape code space gives a natural prior for sampling strokes that fit a part role, which the paper suggests can be reused in other sketch-related tasks.

Reading between the lines

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

  • A step the authors leave implicit is that the discrete code space could support stroke-level editing or retrieval by swapping or interpolating code indices; the paper demonstrates code interpolation for reconstruction but not as a generation-time control.
  • Because the label Transformer can be removed with only a small quality drop, the approach is a plausible fit for unlabeled sketch corpora beyond CreativeSketch, including free-hand doodles without part annotations.
  • A testable extension would be scaling up the codebooks: the paper does not exhaust hyperparameters and reports remaining artifacts like disconnected wings, so larger code spaces might close those gaps.
  • The reported gains are measured in one fixed feature space; using a different visual encoder could reorder the baselines, which would be a useful sensitivity check for the field.
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 / 6 minor

Summary. The paper proposes VQ-SGen, a two-stage generative model for creative sketch drawing. In stage one, each stroke is represented as a triplet of shape, location, and label; the shape and location are separately vector-quantized through a VQ-VAE-style codebook. In stage two, a cascaded autoregressive Transformer generates stroke labels and then shape/location codes, enabling unconditional, class-conditioned, text-conditioned, and completion-based sketch generation. Experiments on CreativeSketch (Creative Birds and Creative Creatures) report improved FID, GD, CS, and SDS over DoodlerGAN, DoodleFormer, and SketchKnitter, together with ablation studies and a user study.

Significance. If the empirical claims hold, VQ-SGen is a useful step for creative sketch generation: it provides a compact discrete stroke representation that captures shape while decoupling position, and it demonstrates that such a representation can support a simple autoregressive generator with several conditioning modes. The paper has several strengths: the two-stage design is clearly described, the ablations in Table 2 directly test the contribution of the VQ representation, decoupling, and label conditioning, and the code-space visualization in Fig. 8 shows semantically meaningful clusters that justify the tokenized formulation. The main limitation is that the central superiority claim rests on point estimates of FID/GD/CS/SDS and on a user study that is partial and unaccompanied by significance testing; the metric validity concern raised below is therefore load-bearing rather than cosmetic.

major comments (4)
  1. [§4.1, Table 1] All quantitative results are reported as single point estimates. The abstract and §4.1 claim that the method 'significantly outperforms' prior work, but without multiple seeds, standard deviations, confidence intervals, or significance tests, the reported margins (FID 15.78 vs 17.48 on Creative Birds and 17.61 vs 20.43 on Creative Creatures) cannot be distinguished from run-to-run variation. Please report mean±std over at least three independent training runs and apply a paired test between the methods.
  2. [§4, 'Evaluation metrics'] The FID, GD, CS, and SDS metrics are all computed in the feature space of an Inception network trained on QuickDraw3.8M. These metrics measure distributional fidelity and diversity in that fixed feature space, not creativity or aesthetic novelty; a baseline that simply returned random training sketches would likely score very well on FID/GD/CS while generating nothing new. The paper itself concedes in §4.1 that a higher CS may reflect 'simplicity of the generation' rather than quality, which undercuts the assumption that the Table 1 metrics are monotonically aligned with the claimed property. Please add a memorization/retrieval baseline, and either include a second metric family (e.g., in a different feature space) or temper the 'surpasses state-of-the-art' claim to 'improves on these metrics in this feature space'.
  3. [§4.3, Fig. 6] The user study compares VQ-SGen only against DoodleFormer and against real dataset sketches, not against DoodlerGAN or SketchKnitter, and no confidence intervals or significance tests are reported. Since the paper concludes from this study that the method 'consistently performs favorably against SoTA methods', the evidence is incomplete. Include all three baselines in the user study, or restrict the claim to the DoodleFormer comparison; in either case, report participant-level statistics.
  4. [§4.1, 'Comparison'] The statement that 'all three methods are trained using their default parameters on our dataset' is underspecified. If the baseline implementations were used with their released defaults, those defaults may have been tuned for other datasets, which could systematically disadvantage the baselines. Please clarify whether any hyperparameter validation was performed for the baselines and, if not, state this as a limitation of the comparison.
minor comments (6)
  1. [§3.1, Eq. (2)] The dimensions of z_i^s and c_j are not given in the main text; please specify them and state the distance metric used for the nearest-code assignment.
  2. [§3.1, Eq. (3)] The balancing weight α multiplies both the codebook and commitment terms; the design choice and the value α=0.8 are reported only in the supplementary. Please state the loss weighting in the main text or move the value to the main description.
  3. [§3.1] The phrase 'build a space for the shape location' reads as a typo; presumably 'shape and location' is intended.
  4. [References] References [1] and [2] are the same SketchINR paper, and [14] and [15] are the same paper; these duplicates should be consolidated.
  5. [§4.3, Fig. 6] The user study figure shows only percentages without the number of participants per comparison or any measure of inter-rater agreement; specifying N and a confidence interval would make the figure interpretable.
  6. [Supplementary D] The DiffSketcher comparison is qualitative and uses a small set of prompts; this is acceptable as a supplementary illustration, but the main text's claim of 'efficient and effective' text-to-sketch generation should not be read as a quantitative comparison.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the representation is trained by reconstruction and VQ losses, the generator by negative log-likelihood, and evaluation uses external metrics and independent baselines.

full rationale

The derivation chain is self-contained. Stage one learns a stroke latent embedding and vector-quantized codebooks through reconstruction, codebook, and commitment losses (Eqs. 1-3), with no evaluation metric used as a training target. Stage two factors the sketch distribution by the chain rule (Eq. 4) and trains cascaded Transformers with a negative log-likelihood loss (Eq. 6), so the generator is not fitted to FID, GD, CS, or SDS. The headline comparison in Table 1 is measured in the feature space of an external Inception model trained on QuickDraw3.8M, against independently trained baselines using default parameters, and the central claim therefore does not reduce to any fitted input. Ablations (w/o VQ, w/o Decouple, w/o T_l) are empirical checks rather than restatements of the method. The observed semantic clustering in the code space is presented as a post-hoc UMAP analysis, not used to fit the model, and the supplementary reports generalization to unseen QuickDraw sketches, which is an external check. The paper does cite the authors' prior ContextSeg for the stroke-entity formulation and for CoordConv/distance-field supervision, but those citations are architectural influences rather than load-bearing evidence for the generation claim, and the relevant design choice is additionally validated by an independent comparison against Sketchformer-based VQ-VAE. No uniqueness theorem, fitted prediction renamed as a result, or self-definitional equivalence appears in the paper. Hence the circularity burden is low and the derivation is not circular.

Assumptions & free parameters 8 free parameters · 6 assumptions · 0 invented entities

All free parameters are standard model hyperparameters selected by validation experiments rather than derived from first principles; they are not fitted to the final evaluation metrics in a circular way. The axioms are background assumptions about dataset validity, metric validity, representation choices from prior work, and the autoregressive factorization.

free parameters (8)
  • Codebook size (V) = 8192
    Number of vectors in both shape and location codebooks; chosen from {2048,4096,8192} via FID/GD on validation (Tables 2 and A1).
  • Code feature dimension = 512
    Dimensionality of each code; 512 chosen over 1024 based on ablation metrics.
  • VQ balancing weight (alpha) = 0.8
    Weight on codebook commitment and reconstruction terms in Eq. 3; set by hand.
  • Maximum stroke length (N) = 20 (Creative Birds), 35 (Creative Creatures)
    Sequence truncation to fit all sketches; affects context range.
  • Learning rates = 1e-4 (VQ), 1e-5 (Gen-Transformer)
    Chosen in supplementary B; no schedule for generator, step decay for VQ.
  • Batch sizes = 64 (VQ), 8 (Gen-Transformer)
    From supplementary B.
  • Scheduled sampling ratio = starts at 1.0, decreases per [20]
    Addresses teacher-forcing gap; exact schedule not specified.
  • Inference sampling threshold (p_n) = not specified numerically
    Nucleus-like sampling keeps top codes with cumulative probability above threshold; value omitted.
assumptions (6)
  • domain assumption The Inception model trained on QuickDraw3.8M provides a valid feature space for evaluating creative sketch quality.
    Used for all quantitative metrics (FID, GD, CS, SDS) in Sec. 4; if this feature space does not capture creative aspects, the rankings may not reflect creativity.
  • domain assumption The CreativeSketch dataset and its part annotations are reliable for training and evaluation.
    The method requires stroke-level part labels for CB/CC; data preprocessing removes 'details' strokes and merges connected strokes (Supplementary A), which changes the data distribution.
  • domain assumption Treating a stroke as an atomic entity and decoupling shape from location is an effective representation for generation.
    Borrowed from ContextSeg [34] and SketchXAI [22]; the paper assumes this carries over from segmentation/explainability to generative modeling.
  • standard math VQ-VAE training with codebook, commitment, and reconstruction losses produces a compact discrete stroke space.
    Standard practice from Van Den Oord et al. [30]; the paper relies on this to make generation tractable.
  • domain assumption Rasterized 256x256 stroke images retain sufficient shape information; distance field supervision improves it.
    The model represents strokes as raster patches rather than point sequences; supplementary compares against Sketchformer and finds raster better.
  • domain assumption The autoregressive factorization p(S)=prod p(v_i^s, v_i^l | l_i) p(l_i) can model coherent sketches.
    The generator (Eq. 4) assumes this conditional structure captures stroke dependencies; if long-range dependencies require a different factorization, generations would degrade.

how reviews work

0 comments
Cite this review

Pith. "Pith review of VQ-SGen: A Vector Quantized Stroke Representation for Creative Sketch Generation." pith.science (2026). https://pith.science/paper/4M2LWGB5

@misc{pith2026241116446,
  author       = {Pith},
  title        = {Pith review of: VQ-SGen: A Vector Quantized Stroke Representation for Creative Sketch Generation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/4M2LWGB5}},
  note         = {Machine review of arXiv:2411.16446}
}
read the original abstract

This paper presents VQ-SGen, a novel algorithm for high-quality creative sketch generation. Recent approaches have framed the task as pixel-based generation either as a whole or part-by-part, neglecting the intrinsic and contextual relationships among individual strokes, such as the shape and spatial positioning of both proximal and distant strokes. To overcome these limitations, we propose treating each stroke within a sketch as an entity and introducing a vector-quantized (VQ) stroke representation for fine-grained sketch generation. Our method follows a two-stage framework - in stage one, we decouple each stroke's shape and location information to ensure the VQ representation prioritizes stroke shape learning. In stage two, we feed the precise and compact representation into an auto-decoding Transformer to incorporate stroke semantics, positions, and shapes into the generation process. By utilizing tokenized stroke representation, our approach generates strokes with high fidelity and facilitates novel applications, such as text or class label conditioned generation and sketch completion. Comprehensive experiments demonstrate our method surpasses existing state-of-the-art techniques on the CreativeSketch dataset, underscoring its effectiveness.

Figures

Figures reproduced from arXiv: 2411.16446 by the authors.

Figure 1
Figure 1. The goal of creative sketch generation is to generate vivid sketches, e.g., the birds from existing methods on the left. We have proposed VQ-SGen, for high-quality sketch generation with a new vector-quantized (VQ) representation and an efficient generator. See one of our results on the right for comparison. places high demands on the generation capabilities of au￾tomated methods. Benefiting from the advancement of … view at source ↗
Figure 2
Figure 2. Overview of VQ-SGen. Given an input sketch, it is first divided into a sequence of strokes. In the first stage, we begin by decoupling the shape and location information of each stroke and obtain their discrete representations (Sec. 3.1). In the second stage, we use a decoder-only Gen-Transformer to predict \left \{ \left ( \bm {I}_i, b_i, l_i \right ) \right \} in an autoregressive manner (Sec. 3.2). pictions of fa… view at source ↗
Figure 3
Figure 3. Network architecture of our generator. on the stroke label. By applying the chain rule, we formu￾late it as: \label {eq:sampling} p(\bm {S}) = \prod _{i=1}^{N} p(\bm {I}_i, b_i, l_i) = \prod _{i=1}^{N} p(v^s_i, v^l_i \mid l_i) \cdot p(l_i). (4) We implement Eq. (4) with two cascaded Transformer decoders, parameterized with θ1 and θ2, respectively. The first decoder samples a label for the next stroke, while the seco… view at source ↗
Figures from the paper (5 more)
Figure 4
Figure 4. Figure 4: Visual comparison against competitors on both the Creative Birds and Creative Creatures datasets. [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 5
Figure 5. Figure 5: Sketch reconstruction results of our ablation study on codebook hyper-parameter setting. The two numbers represent the code [PITH_FULL_IMAGE:figures/full_fig_p006_5.png]
Figure 6
Figure 6. Figure 6: User study results based on the five questions (a-e). [PITH_FULL_IMAGE:figures/full_fig_p007_6.png]
Figure 7
Figure 7. Figure 7: Applications of our approach. (a) Given the class label, our method can produce corresponding sketches with rich variations. [PITH_FULL_IMAGE:figures/full_fig_p008_7.png]
Figure 8
Figure 8. Figure 8: UMAP visualization of the shape codes in [PITH_FULL_IMAGE:figures/full_fig_p008_8.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

45 extracted references · 36 canonical work pages

  1. [1]

    Sketchinr: A first look into sketches as implicit neural representations

    Hmrishav Bandyopadhyay, Ayan Kumar Bhunia, Pinaki Nath Chowdhury, Aneeshan Sain, Tao Xiang, Timothy Hospedales, and Yi-Zhe Song. Sketchinr: A first look into sketches as implicit neural representations. In Proceedings of the IEEE/CVF Conference on Computer Vi- sion and Pattern Recognition (CVPR), pages 12565–12574,

    work page

  2. [2]

    Sketchinr: A first look into sketches as implicit neural representations

    Hmrishav Bandyopadhyay, Ayan Kumar Bhunia, Pinaki Nath Chowdhury, Aneeshan Sain, Tao Xiang, Timothy Hospedales, and Yi-Zhe Song. Sketchinr: A first look into sketches as implicit neural representations. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 12565–12574, 2024. 4

    work page 2024

  3. [3]

    What sketch explainability really means for downstream tasks? In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 10997–11008, 2024

    Hmrishav Bandyopadhyay, Pinaki Nath Chowdhury, Ayan Kumar Bhunia, Aneeshan Sain, Tao Xiang, and Yi-Zhe Song. What sketch explainability really means for downstream tasks? In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 10997–11008, 2024. 1

    work page 2024

  4. [4]

    Doodleformer: Creative sketch drawing with transformers

    Ankan Kumar Bhunia, Salman Khan, Hisham Cholakkal, Rao Muhammad Anwer, Fahad Shahbaz Khan, Jorma Laak- sonen, and Michael Felsberg. Doodleformer: Creative sketch drawing with transformers. In European Conference on Computer Vision, pages 338–355. Springer, 2022. 1, 3, 5, 7

    work page 2022

  5. [5]

    Ai- sketcher: a deep generative model for producing high-quality sketches

    Nan Cao, Xin Yan, Yang Shi, and Chaoran Chen. Ai- sketcher: a deep generative model for producing high-quality sketches. In Proceedings of the AAAI conference on artificial intelligence, pages 2564–2571, 2019. 5

    work page 2019

  6. [6]

    Infogan: Interpretable rep- resentation learning by information maximizing generative adversarial nets

    Xi Chen, Yan Duan, Rein Houthooft, John Schulman, Ilya Sutskever, and Pieter Abbeel. Infogan: Interpretable rep- resentation learning by information maximizing generative adversarial nets. Advances in neural information processing systems, 29, 2016. 2

    2016

  7. [7]

    3d reconstruc- tion from a single sketch via view-dependent depth sampling

    Chenjian Gao, Xilin Wang, Qian Yu, Lu Sheng, Jing Zhang, Xiaoguang Han, Yi-Zhe Song, and Dong Xu. 3d reconstruc- tion from a single sketch via view-dependent depth sampling. IEEE transactions on pattern analysis and machine intelli- gence, 2024. 1

    work page 2024

  8. [8]

    Creativ sketch generation

    Songwei Ge, Vedanuj Goswami, Larry Zitnick, and Devi Parikh. Creativ sketch generation. In International Confer- ence on Learning Representations, 2021. 1, 2, 3, 5, 7

    work page 2021

Show all 45 references
  1. [9]

    A neural representation of sketch drawings

    David Ha and Douglas Eck. A neural representation of sketch drawings. arXiv preprint arXiv:1704.03477 , 2017. 2, 8

    2017 arXiv

  2. [10]

    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. 5

    2017

  3. [11]

    Reducing the dimensionality of data with neural networks

    Geoffrey E Hinton and Ruslan R Salakhutdinov. Reducing the dimensionality of data with neural networks. science, 313(5786):504–507, 2006. 2

    2006

  4. [12]

    Adam: A method for stochastic optimization

    Diederik P Kingma and Jimmy Ba. Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980 ,

    arXiv

  5. [13]

    Text- to-image diffusion models are great sketch-photo matchmak- ers

    Subhadeep Koley, Ayan Kumar Bhunia, Aneeshan Sain, Pinaki Nath Chowdhury, Tao Xiang, and Yi-Zhe Song. Text- to-image diffusion models are great sketch-photo matchmak- ers. In Proceedings of the IEEE/CVF Conference on Com- puter Vision and Pattern Recognition (CVPR), pages 1682...

    2024

  6. [14]

    Subhadeep Koley, Ayan Kumar Bhunia, Aneeshan Sain, Pinaki Nath Chowdhury, Tao Xiang, and Yi-Zhe Song. How to handle sketch-abstraction in sketch-based image re- trieval? In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , pages 16859–1...

    2024

  7. [15]

    How to handle sketch-abstraction in sketch-based image retrieval? In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 16859–16869, 2024

    Subhadeep Koley, Ayan Kumar Bhunia, Aneeshan Sain, Pinaki Nath Chowdhury, Tao Xiang, and Yi-Zhe Song. How to handle sketch-abstraction in sketch-based image retrieval? In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 16859–16869, 2024. 1

    2024

  8. [16]

    Toward deep universal sketch perceptual grouper.IEEE Transactions on Image Pro- cessing, 28(7):3219–3231, 2019

    Ke Li, Kaiyue Pang, Yi-Zhe Song, Tao Xiang, Timothy M Hospedales, and Honggang Zhang. Toward deep universal sketch perceptual grouper.IEEE Transactions on Image Pro- cessing, 28(7):3219–3231, 2019. 2

    2019

  9. [17]

    Fast sketch seg- mentation and labeling with deep learning

    Lei Li, Hongbo Fu, and Chiew-Lan Tai. Fast sketch seg- mentation and labeling with deep learning. IEEE computer graphics and applications, 39(2):38–51, 2018. 2

    2018

  10. [18]

    An intriguing failing of convolutional neural networks and the coordconv solution

    Rosanne Liu, Joel Lehman, Piero Molino, Felipe Pet- roski Such, Eric Frank, Alex Sergeev, and Jason Yosinski. An intriguing failing of convolutional neural networks and the coordconv solution. Advances in neural information pro- cessing systems, 31, 2018. 4

    2018

  11. [19]

    Umap: Uniform manifold approximation and projection for dimen- sion reduction

    Leland McInnes, John Healy, and James Melville. Umap: Uniform manifold approximation and projection for dimen- sion reduction. arXiv preprint arXiv:1802.03426, 2018. 7

    2018 arXiv

  12. [20]

    Sched- uled sampling for transformers

    Tsvetomila Mihaylova and Andr ´e FT Martins. Sched- uled sampling for transformers. arXiv preprint arXiv:1906.07651, 2019. 1

    1906 arXiv

  13. [21]

    Sketchsegnet+: An end- to-end learning of rnn for multi-class sketch semantic seg- mentation

    Yonggang Qi and Zheng-Hua Tan. Sketchsegnet+: An end- to-end learning of rnn for multi-class sketch semantic seg- mentation. Ieee Access, 7:102717–102726, 2019. 2

    2019

  14. [22]

    Sketchxai: A first look at explain- ability for human sketches

    Zhiyu Qu, Yulia Gryaditskaya, Ke Li, Kaiyue Pang, Tao Xi- ang, and Yi-Zhe Song. Sketchxai: A first look at explain- ability for human sketches. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages 23327–23337, 2023. 2

    2023

  15. [23]

    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

  16. [24]

    Sketchformer: Transformer-based representa- tion for sketched structure

    Leo Sampaio Ferraz Ribeiro, Tu Bui, John Collomosse, and Moacir Ponti. Sketchformer: Transformer-based representa- tion for sketched structure. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages 14153–14162, 2020. 2, 4

    2020

  17. [25]

    Stylemeup: Towards style- agnostic sketch-based image retrieval

    Aneeshan Sain, Ayan Kumar Bhunia, Yongxin Yang, Tao Xiang, and Yi-Zhe Song. Stylemeup: Towards style- agnostic sketch-based image retrieval. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 8504–8513, 2021. 1

    2021

  18. [26]

    Freeview sketch- ing: View-aware fine-grained sketch-based image retrieval

    Aneeshan Sain, Pinaki Nath Chowdhury, Subhadeep Koley, Ayan Kumar Bhunia, and Yi-Zhe Song. Freeview sketch- ing: View-aware fine-grained sketch-based image retrieval. In European Conference on Computer Vision , pages 145–

  19. [27]

    Rethinking the inception archi- tecture for computer vision

    Christian Szegedy, Vincent Vanhoucke, Sergey Ioffe, Jon Shlens, and Zbigniew Wojna. Rethinking the inception archi- tecture for computer vision. In Proceedings of the IEEE con- ference on computer vision and pattern recognition , pages 2818–2826, 2016. 5

    2016

  20. [28]

    Strokenuwa: Tokeniz- ing strokes for vector graphic synthesis

    Zecheng Tang, Chenfei Wu, Zekai Zhang, Mingheng Ni, Shengming Yin, Yu Liu, Zhengyuan Yang, Lijuan Wang, Zicheng Liu, Juntao Li, et al. Strokenuwa: Tokeniz- ing strokes for vector graphic synthesis. arXiv preprint arXiv:2401.17093, 2024. 2

    2024 arXiv

  21. [29]

    Sketchdeco: Decorating b&w sketches with colour

    Chaitat Utintu, Pinaki Nath Chowdhury, Aneeshan Sain, Subhadeep Koley, Ayan Kumar Bhunia, and Yi-Zhe Song. Sketchdeco: Decorating b&w sketches with colour. arXiv preprint arXiv:2405.18716, 2024. 1

    2024 arXiv

  22. [30]

    Neural discrete representation learning

    Aaron Van Den Oord, Oriol Vinyals, et al. Neural discrete representation learning. Advances in neural information pro- cessing systems, 30, 2017. 2, 3, 4

    2017

  23. [31]

    Attention is all you need

    A Vaswani. Attention is all you need. Advances in Neural Information Processing Systems, 2017. 3

    2017

  24. [32]

    Stacked denoising autoencoders: Learning useful represen- tations in a deep network with a local denoising criterion

    Pascal Vincent, Hugo Larochelle, Isabelle Lajoie, Yoshua Bengio, Pierre-Antoine Manzagol, and L ´eon Bottou. Stacked denoising autoencoders: Learning useful represen- tations in a deep network with a local denoising criterion. Journal of machine learning research, 11(12), 2010. 2

    2010

  25. [33]

    Multi-column point-cnn for sketch segmentation

    Fei Wang, Shujin Lin, Hanhui Li, Hefeng Wu, Tie Cai, Xi- aonan Luo, and Ruomei Wang. Multi-column point-cnn for sketch segmentation. Neurocomputing, 392:50–59, 2020. 2

    2020

  26. [34]

    Contextseg: Sketch seman- tic segmentation by querying the context with attention

    Jiawei Wang and Changjian Li. Contextseg: Sketch seman- tic segmentation by querying the context with attention. In Proceedings of the IEEE/CVF Conference on Computer Vi- sion and Pattern Recognition, pages 3679–3688, 2024. 1, 2, 4, 7, 3

    2024

  27. [35]

    Sketchknitter: Vectorized sketch generation with dif- fusion models

    Qiang Wang, Haoge Deng, Yonggang Qi, Da Li, and Yi-Zhe Song. Sketchknitter: Vectorized sketch generation with dif- fusion models. In The Eleventh International Conference on Learning Representations, 2023. 1, 5

    2023

  28. [36]

    Sketch- segnet: A rnn model for labeling sketch strokes

    Xingyuan Wu, Yonggang Qi, Jun Liu, and Jie Yang. Sketch- segnet: A rnn model for labeling sketch strokes. In 2018 IEEE 28th International Workshop on Machine Learning for Signal Processing (MLSP), pages 1–6. IEEE, 2018. 2

    2018

  29. [37]

    Diffsketcher: Text guided vector sketch synthesis through latent diffusion models

    Ximing Xing, Chuang Wang, Haitao Zhou, Jing Zhang, Qian Yu, and Dong Xu. Diffsketcher: Text guided vector sketch synthesis through latent diffusion models. Advances in Neu- ral Information Processing Systems, 36:15869–15889, 2023. 1, 2, 3

    2023

  30. [38]

    Deep learning for free-hand sketch: A survey

    Peng Xu, Timothy M Hospedales, Qiyue Yin, Yi-Zhe Song, Tao Xiang, and Liang Wang. Deep learning for free-hand sketch: A survey. IEEE transactions on pattern analysis and machine intelligence, 45(1):285–312, 2022. 5

    2022

  31. [39]

    Sketchgnn: Semantic sketch seg- mentation with graph neural networks

    Lumin Yang, Jiajie Zhuang, Hongbo Fu, Xiangzhi Wei, Kun Zhou, and Youyi Zheng. Sketchgnn: Semantic sketch seg- mentation with graph neural networks. ACM Transactions on Graphics (TOG), 40(3):1–13, 2021. 2

    2021

  32. [40]

    Stro- kenet: A neural painting environment

    Ningyuan Zheng, Yifan Jiang, and Dingjiang Huang. Stro- kenet: A neural painting environment. In International Con- ference on Learning Representations, 2018. 2

    2018

  33. [41]

    Sketch-segformer: Transformer-based segmen- tation for figurative and creative sketches.IEEE Transactions on Image Processing, 2023

    Yixiao Zheng, Jiyang Xie, Aneeshan Sain, Yi-Zhe Song, and Zhanyu Ma. Sketch-segformer: Transformer-based segmen- tation for figurative and creative sketches.IEEE Transactions on Image Processing, 2023. 2

    2023

  34. [42]

    Creativeseg: Semantic seg- mentation of creative sketches

    Yixiao Zheng, Kaiyue Pang, Ayan Das, Dongliang Chang, Yi-Zhe Song, and Zhanyu Ma. Creativeseg: Semantic seg- mentation of creative sketches. IEEE Transactions on Image Processing, 33:2266–2278, 2024. 1

    2024

  35. [43]

    Learning to doodle with stroke demonstrations and deep q-networks

    Tao Zhou, Chen Fang, Zhaowen Wang, Jimei Yang, Byung- moon Kim, Zhili Chen, Jonathan Brandt, and Demetri Ter- zopoulos. Learning to doodle with stroke demonstrations and deep q-networks. In BMVC, page 13, 2018. 2

    2018

  36. [44]

    Part-level sketch segmentation and labeling using dual-cnn

    Xianyi Zhu, Yi Xiao, and Yan Zheng. Part-level sketch segmentation and labeling using dual-cnn. In Neural Infor- mation Processing: 25th International Conference, ICONIP 2018, Siem Reap, Cambodia, December 13-16, 2018, Pro- ceedings, Part I 25, pages 374–384. Springer, 2018. 2

    2018

  37. [45]

    Just Dance

    Xianyi Zhu, Yi Xiao, and Yan Zheng. 2d freehand sketch la- beling using cnn and crf.Multimedia Tools and Applications, 79(1-2):1585–1602, 2020. 2 VQ-SGen: A Vector Quantized Stroke Representation for Creative Sketch Generation Supplementary Material In this supplemental materi...

    2020

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.