Pith. sign in

REVIEW 3 major objections 5 minor 29 references

Spline-FRIDA: Towards Diverse, Humanlike Robot Painting Styles with a Sample-Efficient, Differentiable Brush Stroke Model

T0 review · 3 major / 5 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 This paper claims that representing robot brush strokes as VAE-decoded trajectories with a new differentiable renderer lets a physical robot capture a human's stroke style from fewer than 20 examples, and that human raters judge the…

desk verdict A solid, scoped contribution to robot painting: the new differentiable polyline renderer is the real win, and the style-diversity claim needs explicit scope. read the letter →

arxiv 2412.00597 v2 pith:JHTNI3SG submitted 2024-11-30 cs.RO

Lawrence Chen , Peter Schaldenbrand , Tanmay Shankar , Lia Coleman , Jean Oh This is my paper · ORCID
classification cs.RO
keywords robotpaintingbrushstrokemodelingvariationalautoencoderdifferentiablerenderingstroke-basedtrajectorystyletransfermotioncapturehuman-robotco-creation
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper argues that the limiting factor in robot painting is not the planning algorithm but the stroke primitive itself. It introduces Spline-FRIDA, which represents each brush stroke as a full trajectory sampled from a variational autoencoder trained on motion-captured human drawings, and renders that trajectory with a new differentiable stroke dynamics model, Traj2Stroke. The central claim is that this trajectory representation captures a human artist's intra-stroke style from fewer than 20 demonstration examples, and that paintings planned with it look more human-made, better match the reference image, and are judged more artistic than paintings made with the quadratic Bezier strokes used by FRIDA. A 100-participant survey supports this: participants matched robot drawings to the correct human style source, and preferred Spline-FRIDA over FRIDA on human-likeness, overall quality, reference match, and artistic value. The paper also reports that Traj2Stroke roughly halves the L1 prediction error on out-of-distribution Sharpie strokes compared to a CNN renderer, visibly narrowing the simulation-to-real gap.

What carries the argument

The two load-bearing components are TrajVAE and Traj2Stroke. TrajVAE is a variational autoencoder that compresses a 32-point polyline trajectory, with coordinates $(x,y,h)$, into a 64-dimensional latent code and decodes it back; pretrained on aggregated sessions and fine-tuned on a single drawing, it is the module that encodes stroke style. Traj2Stroke is the differentiable brush stroke dynamics model: it reorients the decoded trajectory to canvas coordinates, then for each segment computes a distance map, a height map by linear interpolation of brush elevation, an affine thickness map $\alpha h + \beta$, and a darkness profile $(1 - d/t)^c$, taking the max over segments. With only seven learnable parameters $m_x, m_y, b_x, b_y, \alpha, \beta, c$, the whole renderer is differentiable end-to-end, so gradient descent can optimize latent vectors $z$, pose offsets $\Delta$, and colors $\rho$ against a CLIP-based feature loss.

What would settle it

Record the same artist drawing the same nominal stroke shape in several different canvas locations and orientations, normalize the trajectories exactly as the paper does, and inspect or train the TrajVAE: if strokes that are identical after normalization produce visibly different real marks, or if the VAE's reconstruction error rises systematically for strokes drawn at certain angles, the normalization premise fails. A more direct test is to define a style by spiral strokes centered at a particular corner of the canvas; since the normalization rotates every trajectory to a horizontal position, such a location-dependent style could not be encoded.

Watch

Extended reading notes

Core claim

The central discovery is that representing a stroke as a polyline trajectory in a learned latent space, rather than as a three-parameter Bezier curve, is what lets a robot adopt a human drawing style. The authors record artists with motion capture, normalize each stroke to a fixed pose, train a variational autoencoder (TrajVAE) on aggregated recordings, and fine-tune it on a single session using fewer than 20 trajectories. During planning the decoder generates stroke trajectories, and a new differentiable renderer, Traj2Stroke, turns each trajectory into a grayscale stroke image with variable thickness using only seven learnable parameters. The paper claims this pipeline preserves distinct human styles, such as long zig-zags, tiny curls, and small circles, and that survey participants reliably identified which human drawing sourced each robot style. In direct comparison, participants judged Spline-FRIDA drawings as more human-like than FRIDA's (73 percent versus 27), better overall (84 versus 16), closer to the reference image (84 versus 16), and more artistic (82 versus 18).

Load-bearing premise

The load-bearing assumption is that the style of a stroke does not depend on where it is drawn on the canvas or at what angle, because each captured trajectory is normalized by moving its start to $(0,0)$ and rotating its endpoint onto the x-axis before the VAE ever sees it.

Editorial extensions

If this is right

  • A user can hand a robot a drawing with a distinctive stroke style, and the robot can reproduce that style in new images after fine-tuning the style autoencoder on fewer than 20 recorded strokes.
  • Because Traj2Stroke is trained once per drawing medium and generalizes to out-of-distribution trajectories, new styles can be added by swapping in a differently fine-tuned TrajVAE without recollecting renderer data.
  • Planning remains fully differentiable: gradient descent can optimize latent stroke codes, canvas poses, and colors directly against a CLIP feature loss, so the method inherits FRIDA's text-, image-, and audio-prompt workflow.
  • The sim-to-real gap for Sharpie strokes is visibly reduced, meaning planned images can be trusted as previews of the physically painted result.
  • This style flexibility comes at a computational cost: a 400-stroke painting takes about an hour to plan on a 4090 GPU, with only 80 strokes optimized at a time due to memory limits.

Reading between the lines

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

  • Because every trajectory is normalized before being encoded, the VAE cannot represent how stroke style depends on canvas position or drawing angle; testing whether artists actually vary stroke shape with pose would directly extend the method.
  • Traj2Stroke renders each segment independently and composites by max darkness, so it cannot model how overlapping wet strokes physically mix; a sequential renderer with ink transfer is a natural next step.
  • The survey compares the new trajectory representation against one Bezier-based system, so how much of the gain comes from the renderer versus the representation itself is untested and could be probed by giving the Bezier planner the same renderer.
  • The paper limits itself to intra-stroke style, leaving inter-stroke elements such as spacing, stroke order, and overlap unmodeled; a hierarchical stroke-pattern model could capture those without changing Traj2Stroke.
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 / 5 minor

Summary. The paper introduces Spline-FRIDA, a robotic painting system that models human brush stroke trajectories using a variational autoencoder (TrajVAE) and a new differentiable renderer (Traj2Stroke). The authors collect motion-capture data of human drawings, fine-tune per-style TrajVAEs from fewer than 20 examples, and plan paintings via gradient descent through Traj2Stroke. Human evaluations on Mechanical Turk suggest that Spline-FRIDA captures stroke styles and produces more human-like, higher-quality, and more artistic drawings than the Bézier-based FRIDA baseline. The renderer is evaluated on out-of-distribution trajectories against CNN baselines, with Traj2Stroke achieving the lowest sharpie loss and Traj2Stroke with U-Net the lowest brush loss.

Significance. If the results hold, Spline-FRIDA offers a practical path to transferring a user's stroke style to a physical robot from a small number of demonstrations, and Traj2Stroke provides a sample-efficient differentiable renderer with a small Sim2Real gap. The paper's strengths include the Real2Sim2Real data collection, the modular VAE-decoder latent space, the out-of-distribution renderer experiments, and the use of physical robot executions in the human evaluation. The authors also honestly acknowledge computational limitations and the normalization tradeoff. However, the central diversity and human-likeness claims are broader than the current evidence, mainly because one load-bearing modeling assumption is untested and the survey results lack inferential statistics.

major comments (3)
  1. [V-B, VI] The trajectory normalization described in Section V-B translates every captured stroke so its start is at (0,0) and rotates it so its endpoint lies on the x-axis. The authors explicitly note this 'assumes trajectory style is not affected by position/rotation on the canvas,' but they do not test this assumption or limit the paper's claims to rotation-invariant styles. As a result, the TrajVAE cannot encode orientation-dependent stroke styles (e.g., hatching aligned with a form, slanted handwriting), so the claimed ability to model 'diverse human brush stroke trajectories' (abstract) and the style-transfer results in Section VI rest on an untested invariance. I would like to see either an experiment that checks VAE reconstruction on rotated/translated versions of held-out strokes and planning with different orientations, or a revised claim that explicitly restricts the method to styles that are position/rotation-invariant.
  2. [VI-A, Table I] The human evaluation in Section VI-A reports raw participant counts for five binary questions but provides no significance tests, confidence intervals, or inter-rater agreement measures. For example, the abstract's claim that Spline-FRIDA strokes are 'more human-like' is based on a 73/27 split in Table I; a binomial test would presumably support this, but no test is shown, and for the more balanced 'abstract' question (40/60) no conclusion can be drawn from the reported data alone. The claim that Spline-FRIDA 'improve[s] semantic planning' is supported only by the subjective 'better matches the reference image' question, not by a direct measure of planning objective loss. Please add appropriate statistical analyses and either directly measure planning quality (e.g., final loss or image similarity) or rephrase the claim.
  3. [VI-C, Table II] The conclusion that Traj2Stroke has a 'significantly smaller Sim2Real gap' (Section I contribution list) is based on average L1 losses in Table II, but the table reports no variance, error bars, or test of significance across repeated runs or trajectories. While the sharpie margin (0.00055 vs 0.00095) looks meaningful, the manuscript does not establish reliability, and on brush strokes Traj2Stroke with U-Net is the best method, which qualifies the headline result. Please report standard deviations or confidence intervals over the evaluation set and clarify that the choice of base Traj2Stroke for the full system is a design decision driven by generalization and simplicity rather than by absolute loss on the brush medium.
minor comments (5)
  1. [V-C] The training objective of TrajVAE is described as minimizing mean squared error; as a variational autoencoder it should also include the KL divergence term. Please state the full loss or clarify that only the reconstruction term is shown.
  2. [VI] The five human drawings used in Fig. 6 are hand-picked lab members' drawings; while acceptable for a proof-of-concept, the paper would benefit from stating that the styles were selected and from reporting the number of trajectories and artists per style.
  3. [VII] There is a typo in 'inneficiencies' (should be 'inefficiencies') and a few other small grammatical errors that should be corrected in a final revision.
  4. [V-D, Eq. (1)] If a trajectory has a zero-length segment (v = u), Eqs. (1)-(3) involve division by zero; please state the assumed handling for degenerate segments or add a small epsilon.
  5. [VI-C] The qualitative claim that 'the vanilla CNN with and without CoordConv fails to generalize' is supported by Fig. 9, but the quantitative Table II does not break down the loss per style; reporting per-style losses would make the generalization claim more precise.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the style-capture and rendering claims are supported by external human evaluation and out-of-distribution testing, not by fitted inputs or self-citation chains.

full rationale

The paper's derivation chain is not circular. TrajVAE is trained on motion-captured human stroke trajectories and evaluated by a human matching task in which raters identify which robot drawing matches a given human drawing's stroke style; this is a direct fidelity test of the generative model against external human judgment, not a fitted quantity renamed as a prediction. Traj2Stroke's seven learnable parameters are fit on robot-executed strokes and then evaluated quantitatively and qualitatively on dataset B, which consists of trajectories from TrajVAEs not used in training, so the generalization claim is tested out-of-distribution rather than assumed. The comparison to FRIDA uses the authors' prior system as a baseline, but the human evaluation is external and the conclusion does not rely on FRIDA's correctness; it is a comparative result. The acknowledged normalization in Section V-B (translating start to origin and rotating endpoint onto the x-axis) is an untested modeling assumption about rotation/position invariance, not a circular step: it does not define the target result in terms of the input or fit the target claim into the preprocessing. No self-citation is load-bearing in the sense of importing an unverified uniqueness or ansatz, and no known result is merely renamed. Overall, the central claims have independent empirical support and do not reduce to their own inputs.

Assumptions & free parameters 5 free parameters · 6 assumptions · 0 invented entities

The system has two learned components: the TrajVAE trajectory model and the seven Traj2Stroke parameters. The rendering equations are hand-specified rule-based forms with constants fit to data; none are derived from physics. The paper introduces no new entities beyond learned parameters and does not claim an external falsifiable object.

free parameters (5)
  • Traj2Stroke affine calibration (mx, my, bx, by) = learned
    Section V-D: maps trajectory coordinates to canvas coordinates after rotation; learned from before/after stroke photos to correct camera calibration error.
  • Height-to-thickness slope alpha and intercept beta = learned
    Section V-D Eq. 4: Thickness = alpha * Height + beta; fitted to stroke renderings.
  • Darkness dropoff exponent c = learned
    Section V-D Eq. 5: darkness = clamp(1 - distance/thickness)^c; fitted to data.
  • TrajVAE encoder/decoder weights = trained
    Section V-C: mean squared error reconstruction training on mocap trajectories; central to style modeling.
  • Per-stroke height parameters during planning = optimized per painting
    Section V-C: height is not encoded by TrajVAE; it is optimized as separate stroke parameters during planning, so style output depends on optimization.
assumptions (6)
  • domain assumption Stroke style is invariant to translation and rotation on the canvas
    Section V-B: all trajectories are translated and rotated to a canonical pose before encoding; explicitly acknowledged tradeoff.
  • domain assumption Brush tip height and stroke thickness are affinely related
    Section V-D Eq. 4: thickness = alpha*height + beta with learned constants; no physical model justifies affine form.
  • ad hoc to paper Darkness falls off as (1 - d/thickness)^c from the stroke center
    Section V-D Eq. 5: chosen to model gradual dropoff; exponent c is fit, the functional form is assumed.
  • domain assumption Linear interpolation of height along a segment approximates real stroke appearance
    Section V-D Eqs. 2-3: height map linearly interpolates between h_k and h_{k+1}.
  • domain assumption Pretrained CLIP features are a suitable loss for matching paintings to target images
    Section V-F: planning uses cosine similarity in CLIP feature space, inherited from prior work; no validation within this paper.
  • domain assumption TrajVAE latent space is smooth enough for gradient-based optimization of z
    Section VI-B uses t-SNE clusters as evidence; planning differentiates through the decoder and optimizes z.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Spline-FRIDA: Towards Diverse, Humanlike Robot Painting Styles with a Sample-Efficient, Differentiable Brush Stroke Model." pith.science (2026). https://pith.science/paper/JHTNI3SG

@misc{pith2026241200597,
  author       = {Pith},
  title        = {Pith review of: Spline-FRIDA: Towards Diverse, Humanlike Robot Painting Styles with a Sample-Efficient, Differentiable Brush Stroke Model},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/JHTNI3SG}},
  note         = {Machine review of arXiv:2412.00597}
}
read the original abstract

A painting is more than just a picture on a wall; a painting is a process comprised of many intentional brush strokes, the shapes of which are an important component of a painting's overall style and message. Prior work in modeling brush stroke trajectories either does not work with real-world robotics or is not flexible enough to capture the complexity of human-made brush strokes. In this work, we introduce Spline-FRIDA which can model complex human brush stroke trajectories. This is achieved by recording artists drawing using motion capture, modeling the extracted trajectories with an autoencoder, and introducing a novel brush stroke dynamics model to the existing robotic painting platform FRIDA. We conducted a survey and found that our open-source Spline-FRIDA approach successfully captures the stroke styles in human drawings and that Spline-FRIDA's brush strokes are more human-like, improve semantic planning, and are more artistic compared to existing robot painting systems with restrictive B\'ezier curve strokes.

Figures

Figures reproduced from arXiv: 2412.00597 by the authors.

Figure 1
Figure 1. Spline-FRIDA drawings and paintings in different styles. These are two pairs of artworks of made by our system. The left paintings use longer, zig-zagging strokes, while the right ones are composed of small circles and dots. While each pair depicts the same content, the stroke style vastly changes the appearance and vibe of each work. over the style of the image, allows them to feel more ownership over the artwork t… view at source ↗
Figure 2
Figure 2. Weaknesses in FRIDA’s stroke model. Left: FRIDA struggles with rendering thin strokes. Right: for a full drawing, errors in individual stroke predictions accumulate. The optimization process exploits blotchy predictions made by the renderer to create small dots, which appear quite different when actually executed. This is mainly because differentiable environment allows for end-to-end training of the RL agent. Paint… view at source ↗
Figure 3
Figure 3. Traj2Stroke. The inputs are a latent vector z and an offset ∆. z is fed through the decoder of a TrajVAE, generating a raw trajectory, which is then rotated and translated according to ∆. We then process the trajectory segments independently, obtaining darkness values for each. Finally, we take the max darkness over all segments. (0,0) x y Translate & Rotate [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗
Figures from the paper (6 more)
Figure 4
Figure 4. Figure 4: Mocap setup. We use a motion capture system to track the position of the canvas and pen over time as an artist draws. Three mocap markers are placed along the corners of the canvas, and four are mounted at the end of the pen. The trajectories of each stroke are extract…
Figure 5
Figure 5. Figure 5: Planning a Painting. As described in Section V-F, Spline-FRIDA plans a painting by optimizing the brush stroke parameters through the dynamics model to decrease a features space loss between a given image and the planned painting. Whereas FRIDA models brush strokes as …
Figure 6
Figure 6. Figure 6: Example drawings made by Spline-FRIDA. Each column represents a distinct trajectory style and each row uses a different objective. The top row contains original drawings made by human artists on our mocap system. One VAE was fine-tuned on each human drawing and used to…
Figure 7
Figure 7. Figure 7: Confusion matrix for matching task. The x-axis represents the index of the specific TrajVAE used to generate the drawing, and the y-axis represents the index of the human drawing participants thought was most similar. The five human drawings/styles the same ones as in …
Figure 9
Figure 9. Figure 9: Visualizing the outputs of various stroke models. The first three rows contain sharpie strokes, and the last three contain brush strokes all of which were made from samples using TrajVAE models not used for training. meaning that these trajectories are out of distribut…
Figure 10
Figure 10. Figure 10: Spline-FRIDA’s low Sim2Real gap. We compare a plan made by Spline-FRIDA with its execution (physically drawn with a robot). The top row with a black marker, and the bottom row with a paint brush. improve runtime planning speeds, memory constraints, and potentially hel…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

29 extracted references · 25 canonical work pages

  1. [1]

    The dichotomy of form and content,

    Duncan Robertson, “The dichotomy of form and content,” College English, vol. 28, no. 4, pp. 273–279, 1967

    work page 1967

  2. [2]

    Styleclipdraw: Coupling content and style in text-to-drawing translation,

    Peter Schaldenbrand, Zhixuan Liu, and Jean Oh, “Styleclipdraw: Coupling content and style in text-to-drawing translation,”arXiv preprint arXiv:2202.12362, 2022

    arXiv 2022

  3. [3]

    Ai art and its impact on artists,

    Harry H. Jiang, Lauren Brown, Jessica Cheng, Mehtab Khan, Abhishek Gupta, Deja Workman, Alex Hanna, Johnathan Flowers, and Timnit Gebru, “Ai art and its impact on artists,” in Proceedings of the 2023 AAAI/ACM Conference on AI, Ethics, and Society , New York, NY , USA, 2023, AIES ’23, p. 363–374, Association for Computing Machinery

    work page 2023

  4. [4]

    Exploring co-creative drawing workflows,

    Chipp Jansen and Elizabeth Sklar, “Exploring co-creative drawing workflows,” Frontiers in Robotics and AI , vol. 8, pp. 577770, 2021

    work page 2021

  5. [5]

    thesis, Harvard University, 2021

    Cole Bateman, Creating for Creatives: A Humanistic Approach to Designing AI Tools Targeted at Professional Animators , Ph.D. thesis, Harvard University, 2021

    work page 2021

  6. [6]

    Empirically studying participatory sense-making in abstract drawing with a co-creative cognitive agent,

    Nicholas Davis, Chih-PIn Hsiao, Kunwar Yashraj Singh, Lisa Li, and Brian Magerko, “Empirically studying participatory sense-making in abstract drawing with a co-creative cognitive agent,” in Proceedings of the 21st International Conference on Intelligent User Interfaces , 2016, pp. 196–207

    work page 2016

  7. [7]

    Drawing with reframer: Emergence and control in co-creative ai,

    Tomas Lawton, Francisco J Ibarrola, Dan Ventura, and Kazjon Grace, “Drawing with reframer: Emergence and control in co-creative ai,” in Proceedings of the 28th International Conference on Intelligent User Interfaces, 2023, pp. 264–277

    work page 2023

  8. [8]

    Adversarial Robots as Creative Collaborators

    Shayla Lee and Wendy Ju, “Adversarial robots as creative collaborators,” arXiv preprint arXiv:2402.03691 , 2024

    work page Pith review arXiv 2024

Show all 29 references
  1. [9]

    Learning to paint with model-based deep reinforcement learning,

    Zhewei Huang, Wen Heng, and Shuchang Zhou, “Learning to paint with model-based deep reinforcement learning,” 2019

    2019

  2. [10]

    Frida: A col- laborative robot painter with a differentiable, real2sim2real planning environment,

    Peter Schaldenbrand, James McCann, and Jean Oh, “Frida: A col- laborative robot painter with a differentiable, real2sim2real planning environment,” 2022

    2022

  3. [11]

    Stylized neural painting,

    Zhengxia Zou, Tianyang Shi, Shuang Qiu, Yi Yuan, and Zhenwei Shi, “Stylized neural painting,” 2020

    2020

  4. [12]

    Content masked loss: Human-like brush stroke planning in a reinforcement learning painting agent,

    Peter Schaldenbrand and Jean Oh, “Content masked loss: Human-like brush stroke planning in a reinforcement learning painting agent,” 2021

    2021

  5. [13]

    Paint transformer: Feed forward neural painting with stroke prediction,

    Songhua Liu, Tianwei Lin, Dongliang He, Fu Li, Ruifeng Deng, Xin Li, Errui Ding, and Hao Wang, “Paint transformer: Feed forward neural painting with stroke prediction,” 2021

    2021

  6. [14]

    Towards layer-wise image vectorization,

    Xu Ma, Yuqian Zhou, Xingqian Xu, Bin Sun, Valerii Filev, Nikita Orlov, Yun Fu, and Humphrey Shi, “Towards layer-wise image vectorization,” 2022

    2022

  7. [15]

    Help- inghand: example-based stroke stylization,

    Jingwan Lu, Fisher Yu, Adam Finkelstein, and Stephen DiVerdi, “Help- inghand: example-based stroke stylization,” ACM Trans. Graph. , vol. 31, no. 4, jul 2012

    2012

  8. [16]

    My text in your handwriting,

    Tom S. F. Haines, Oisin Mac Aodha, and Gabriel J. Brostow, “My text in your handwriting,” ACM Trans. Graph. , vol. 35, no. 3, may 2016

    2016

  9. [17]

    Differentiable vector graphics rasterization for editing and learning,

    Tzu-Mao Li, Michal Luk ´aˇc, Gharbi Micha ¨el, and Jonathan Ragan- Kelley, “Differentiable vector graphics rasterization for editing and learning,” ACM Trans. Graph. (Proc. SIGGRAPH Asia) , vol. 39, no. 6, pp. 193:1–193:15, 2020

    2020

  10. [18]

    Vectorfusion: Text-to-svg by abstracting pixel-based diffusion models,

    Ajay Jain, Amber Xie, and Pieter Abbeel, “Vectorfusion: Text-to-svg by abstracting pixel-based diffusion models,” 2022

    2022

  11. [19]

    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,” in Advances in Neural Information Processing Systems, A. Oh, T. Naumann, A. Globerson, K. Saenko, M. Hardt, and S. Levi...

    2023

  12. [20]

    Svgdreamer: Text guided svg generation with diffusion model,

    Ximing Xing, Haitao Zhou, Chuang Wang, Jing Zhang, Dong Xu, and Qian Yu, “Svgdreamer: Text guided svg generation with diffusion model,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , June 2024, pp. 4546–4555

    2024

  13. [21]

    Robot synesthesia: A sound and emotion guided ai painter,

    Vihaan Misra, Peter Schaldenbrand, and Jean Oh, “Robot synesthesia: A sound and emotion guided ai painter,” 2023

    2023

  14. [22]

    Cofrida: Self-supervised fine-tuning for human-robot co-painting,

    Peter Schaldenbrand, Gaurav Parmar, Jun-Yan Zhu, James McCann, and Jean Oh, “Cofrida: Self-supervised fine-tuning for human-robot co-painting,” in 2024 IEEE International Conference on Robotics and Automation (ICRA) . IEEE, 2024

    2024

  15. [23]

    Auto-encoding variational bayes,

    Diederik P Kingma and Max Welling, “Auto-encoding variational bayes,” 2022

    2022

  16. [24]

    Learning trans- ferable visual models from natural language supervision,

    Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, Gretchen Krueger, and Ilya Sutskever, “Learning trans- ferable visual models from natural language supervision,” CoRR, vol. abs/2...

    2021 arXiv

  17. [25]

    Clipasso: Semantically-aware object sketching,

    Yael Vinker, Ehsan Pajouheshgar, Jessica Y Bo, Roman Christian Bachmann, Amit Haim Bermano, Daniel Cohen-Or, Amir Zamir, and Ariel Shamir, “Clipasso: Semantically-aware object sketching,” arXiv preprint arXiv:2202.05822, 2022

    2022 arXiv

  18. [26]

    Visualizing data using t-SNE,

    Laurens van der Maaten and Geoffrey Hinton, “Visualizing data using t-SNE,” Journal of Machine Learning Research , vol. 9, pp. 2579–2605, 2008

    2008

  19. [27]

    An intriguing failing of convolutional neural networks and the coordconv solution,

    Rosanne Liu, Joel Lehman, Piero Molino, Felipe Petroski Such, Eric Frank, Alex Sergeev, and Jason Yosinski, “An intriguing failing of convolutional neural networks and the coordconv solution,” 2018

    2018

  20. [28]

    U-net: Convo- lutional networks for biomedical image segmentation,

    Olaf Ronneberger, Philipp Fischer, and Thomas Brox, “U-net: Convo- lutional networks for biomedical image segmentation,” 2015. 9

    2015

  21. [29]

    Adding conditional control to text-to-image diffusion models,

    Lvmin Zhang, Anyi Rao, and Maneesh Agrawala, “Adding conditional control to text-to-image diffusion models,” 2023

    2023

Pith tools

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