pith. machine review for the scientific record. sign in

arxiv: 2604.06489 · v1 · submitted 2026-04-07 · 💻 cs.HC · cs.MM

Recognition: 2 theorem links

· Lean Theorem

Language-Guided Multimodal Texture Authoring via Generative Models

Aiden Chang, Heather Culbertson, Michael Gu, Shihan Lu, Wanli Qian

Authors on Pith no claims yet

Pith reviewed 2026-05-10 18:31 UTC · model grok-4.3

classification 💻 cs.HC cs.MM
keywords haptic texture authoringlanguage-guided generationmultimodal texturesgenerative modelscross-modal consistencydiffusion modelsautoregressive haptic models
0
0 comments X

The pith

A shared language-aligned latent space lets single text prompts generate coordinated haptic vibrations, tapping transients, and visual previews for texture authoring.

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

The paper shows how natural-language descriptions can replace manual parameter tuning when creating textures that users both see and feel. Generative models for sliding vibrations and tapping sounds are aligned with a diffusion model for images through a common latent representation trained to respect language semantics. Participant ratings of generated samples, anchored to real materials, project onto consistent perceptual lines for roughness, slipperiness, and hardness, indicating that the latent structure captures meaningful material attributes. A user study finds that the resulting system supports low-effort, prompt-based iteration while producing coherent cross-modal experiences.

Core claim

By training a shared latent space that aligns language with multiple generative models, one prompt produces semantically matched outputs: force- and speed-conditioned autoregressive sequences for sliding haptic vibrations, transient signals for tapping, and diffusion-based visual renderings. Anchor-referenced ratings reveal that the outputs track expected perceptual trends, such as asperity effects in roughness and compliance in hardness, while participants report coherent multimodal sensations and reduced authoring effort.

What carries the argument

The shared, language-aligned latent space that links text prompts to coordinated outputs from autoregressive haptic models and diffusion visual models.

If this is right

  • Designers can author multimodal textures by writing goals in plain language rather than adjusting low-level parameters through trial and error.
  • A single prompt simultaneously steers sliding vibrations, tapping transients, and visual appearance while preserving semantic consistency.
  • Ratings projected between real references demonstrate that roughness tracks surface asperity, hardness tracks compliance, and slipperiness tracks surface film effects.
  • Users experience coherent cross-modal feedback and report low effort when iterating via text prompts on a 3D haptic device.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same alignment technique could be tested on additional attributes such as temperature or stickiness to expand the range of controllable material properties.
  • Integration with real-time rendering engines might allow prompt-driven texture updates during interactive design sessions.
  • The approach suggests a path for extending language control to other paired modalities, such as sound and force profiles, if similar latent alignment can be achieved.

Load-bearing premise

The learned latent space captures perceptually meaningful structure that generalizes beyond the training examples and the three evaluated material attributes.

What would settle it

Participant ratings of the generated textures fail to form consistent, interpretable trends when projected onto lines between real-material anchors for roughness, slipperiness, or hardness, or different prompts produce outputs that participants judge as semantically mismatched across the haptic and visual channels.

Figures

Figures reproduced from arXiv: 2604.06489 by Aiden Chang, Heather Culbertson, Michael Gu, Shihan Lu, Wanli Qian.

Figure 1
Figure 1. Figure 1: System overview. A text prompt is encoded into a haptic latent z. Modality-specific decoders then output (i) a tap bank for hardness rendering and (ii) an AR matrix for sliding-produced vibrations. The same prompt also drives a decoupled text-to-image diffusion model for the visual preview. Here, y[n] is the vibration signal at n th time step and ak(f, v) are the AR coefficients corresponding to the curren… view at source ↗
Figure 2
Figure 2. Figure 2: Learning latent. A bimodal VAE reconstructs tap bank and AR matrix from haptic latent z; a contrastive head aligns z with CLIP text, with KL regularization keeping the latent compact and reconstruction losses preserving haptic fidelity. withholding distinct textures would create large semantic gaps in the latent space, severely impairing the model’s ability to extrapolate or interpolate between material ca… view at source ↗
Figure 3
Figure 3. Figure 3: VAE latent space embedding. Each point represents the posterior mean µ for one of the 100 texture classes after dimensionality reduction (PCA + t-SNE). Nearby points correspond to materials that the model internally considers perceptually similar. IV. EXPERIMENTAL RESULTS We assess whether the learned haptic latent (trained on AR matrix + tap bank only) captures perceptually consistent relationships among … view at source ↗
Figure 4
Figure 4. Figure 4: Tri-modal texture generation from language. [PITH_FULL_IMAGE:figures/full_fig_p007_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Latent interpolation exemplar. Columns show TOWEL→latent midpoint→WHITEBOARD. Top: diffusion image (semantics only, no imposed orientation). Middle: synthesized tap responses across 13 impact speeds (higher saturation = higher speed). Bottom: AR pole distributions on the unit circle, colored by excitation variance. Given the posterior means µA and µB for two anchors, we compute their average z¯ = 1 2 (µA +… view at source ↗
Figure 6
Figure 6. Figure 6: User study setup. Left: Participant interacting with the 3D Systems Touch device while viewing the texture and rating UI. Right: On-screen interface showing the rendered texture, controls, and three sliders for rating Roughness, Slipperiness, and Hardness. Hooks (selected from HaTT Corpus for distinct perceptual features). For each pair, we generated the haptic intermediate by decoding the arithmetic mean … view at source ↗
Figure 8
Figure 8. Figure 8: Attribute-wise axis projections with distribution. Violin plots (density) overlaid with per-trial scatter. The green band marks “between anchors” (ta∈ [0, 1]). We grounded our questionnaire in the Haptic Experience Inventory (HXI) factors—Autotelics, Involvement, Realism, Harmony, Discord (4 items each)—and adapted items to our usecase following [81]. Items used a 7-point Likert scale [PITH_FULL_IMAGE:fig… view at source ↗
Figure 10
Figure 10. Figure 10: Interaction usefulness for Sliding, Tapping, their combination, and 3D visualization. Distributions concentrate toward the satisfactory (blue) end; the combination shifts fur￾thest right, indicating complementarity rather than redundancy in haptic channels [PITH_FULL_IMAGE:figures/full_fig_p010_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: Attribute-rating ease. Roughness is easiest, Hard￾ness next, Slipperiness hardest—consistent with our signal pathways (roughness salient while sliding; hardness from tap transients; slipperiness depends on friction/adhesion cues) that some users find subtler to interpret. Notably, the Slid￾ing+Tapping condition shifts the entire distribution further to￾ward satisfaction, showing the two actions are comple… view at source ↗
Figure 12
Figure 12. Figure 12: NASA–TLX Mental, physical, temporal demand, effort, frustration, and perceived success. Scores are shown on their native scales; lower is better for demand/effort/frus￾tration, higher is better for success [PITH_FULL_IMAGE:figures/full_fig_p011_12.png] view at source ↗
Figure 13
Figure 13. Figure 13: Language-specific outcomes. Agreement that the output matched the description and felt realistic, plus self￾rated ability to create and refine using language prompts. In [PITH_FULL_IMAGE:figures/full_fig_p011_13.png] view at source ↗
read the original abstract

Authoring realistic haptic textures typically requires low-level parameter tuning and repeated trial-and-error, limiting speed, transparency, and creative reach. We present a language-driven authoring system that turns natural-language prompts into multimodal textures: two coordinated haptic channels - sliding vibrations via force/speed-conditioned autoregressive (AR) models and tapping transients - and a text-prompted visual preview from a diffusion model. A shared, language-aligned latent links modalities so a single prompt yields semantically consistent haptic and visual signals; designers can write goals (e.g., "gritty but cushioned surface," "smooth and hard metal surface") and immediately see and feel the result through a 3D haptic device. To verify that the learned latent encodes perceptually meaningful structure, we conduct an anchor-referenced, attribute-wise evaluation for roughness, slipperiness, and hardness. Participant ratings are projected to the interpretable line between two real-material references, revealing consistent trends - asperity effects in roughness, compliance in hardness, and surface-film influence in slipperiness. A human-subject study further indicates coherent cross-modal experience and low effort for prompt-based iteration. The results show that language can serve as a practical control modality for texture authoring: prompts reliably steer material semantics across haptic and visual channels, enabling a prompt-first, designer-oriented workflow that replaces manual parameter tuning with interpretable, text-guided refinement.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

3 major / 2 minor

Summary. The paper presents a language-guided multimodal texture authoring system that maps natural-language prompts to coordinated haptic outputs (force/speed-conditioned AR models for sliding vibrations and tapping transients) and visual previews (diffusion model) via a shared language-aligned latent space. A single prompt is claimed to produce semantically consistent cross-modal results. Verification relies on an anchor-referenced user study in which participant ratings for roughness, slipperiness, and hardness are projected onto lines between real-material references, yielding 'consistent trends,' plus a separate human-subject study reporting coherent cross-modal experience and low iteration effort.

Significance. If the central claim holds, the work could meaningfully advance prompt-based, designer-oriented workflows in haptics and HCI by replacing manual parameter tuning with interpretable language control. The anchor-referenced projection method and cross-modal consistency demonstration are potentially useful contributions. However, the absence of quantitative metrics, model details, ablations, and controls for visual bias limits the strength of the evidence that the latent encodes generalizable perceptual structure.

major comments (3)
  1. [User evaluation / anchor-referenced study] User evaluation section: the anchor-referenced projection of ratings onto lines between real-material references for roughness, slipperiness, and hardness implicitly assumes linear perceptual interpolation and that the chosen anchors span the relevant semantics without bias. No validation of linearity, no ablation of anchor choice, and no controls separating haptic signals from the accompanying visual diffusion preview are provided; if ratings are driven primarily by visuals, the evidence that the shared latent itself encodes perceptually meaningful cross-modal structure collapses.
  2. [Methods / system architecture] Methods / system description: no quantitative metrics, training details, model architectures, loss functions, or error analysis are reported for the AR haptic models, diffusion visual model, or the shared language-aligned latent. Without these, the reliability of the semantic-steering claim cannot be assessed and the work is not reproducible.
  3. [Evaluation / experiments] No ablation studies (e.g., shared latent vs. independent modality models or language vs. non-language conditioning) are described. Such comparisons are load-bearing for the claim that the language-aligned latent is what produces the observed cross-modal consistency.
minor comments (2)
  1. [Abstract / Introduction] The abstract and introduction could more clearly distinguish the novel contribution (shared latent) from the component generative models.
  2. [Figures / evaluation plots] Figure captions and axis labels in the evaluation plots should explicitly state the projection method and any statistical tests used to support 'consistent trends.'

Simulated Author's Rebuttal

3 responses · 2 unresolved

We thank the referee for the constructive and detailed feedback. The comments identify key areas where additional clarification, details, and caveats can strengthen the manuscript. We address each major comment below and indicate the revisions we will make.

read point-by-point responses

  1. Referee: User evaluation section: the anchor-referenced projection of ratings onto lines between real-material references for roughness, slipperiness, and hardness implicitly assumes linear perceptual interpolation and that the chosen anchors span the relevant semantics without bias. No validation of linearity, no ablation of anchor choice, and no controls separating haptic signals from the accompanying visual diffusion preview are provided; if ratings are driven primarily by visuals, the evidence that the shared latent itself encodes perceptually meaningful cross-modal structure collapses.

    Authors: We agree that the anchor-referenced projection method rests on an unvalidated linearity assumption and that the study did not ablate anchor selection or isolate haptic ratings from visual previews. These are legitimate concerns that could affect interpretation of the latent's cross-modal contribution. In the revised manuscript we will add an explicit limitations subsection in the evaluation that states these assumptions, notes the lack of linearity validation and modality-isolation controls, and discusses potential visual bias. We will also emphasize that the observed attribute trends align with established haptic literature. We cannot perform new controlled experiments for this revision but will flag the need for such studies as future work. revision: partial

  2. Referee: Methods / system description: no quantitative metrics, training details, model architectures, loss functions, or error analysis are reported for the AR haptic models, diffusion visual model, or the shared language-aligned latent. Without these, the reliability of the semantic-steering claim cannot be assessed and the work is not reproducible.

    Authors: We acknowledge that the current manuscript provides insufficient technical detail on the generative components, limiting reproducibility and assessment of the semantic alignment claim. The focus was on the integrated authoring workflow and user evaluation. In the revision we will add an appendix containing the AR haptic model architecture and conditioning, diffusion model configuration, the shared latent alignment procedure, loss functions employed, training hyperparameters, and any available quantitative metrics or error statistics from our experiments. revision: yes

  3. Referee: No ablation studies (e.g., shared latent vs. independent modality models or language vs. non-language conditioning) are described. Such comparisons are load-bearing for the claim that the language-aligned latent is what produces the observed cross-modal consistency.

    Authors: We concur that ablation studies comparing the shared latent against independent modality models or language versus non-language conditioning would provide stronger evidence that the language-aligned latent drives the observed consistency. These experiments were not conducted in the original work. In the revised manuscript we will add a dedicated limitations paragraph that explicitly notes the absence of such ablations, explains their importance for validating the latent's role, and identifies them as a priority for future research. We cannot design, run, and report new ablation experiments within the scope of this revision. revision: no

standing simulated objections not resolved
  • New controlled user studies with modality separation and linearity validation
  • Ablation experiments on the shared latent versus independent models

Circularity Check

0 steps flagged

No significant circularity; evaluation relies on independent participant data

full rationale

The paper describes a generative system with a shared language-aligned latent for consistent multimodal textures and verifies perceptual meaningfulness via an anchor-referenced user study. Participant ratings on generated outputs are projected onto lines between real-material references to show trends in roughness, slipperiness, and hardness. This constitutes external empirical validation using human judgments independent of model training or fitting. No equations, self-definitional mappings, fitted parameters renamed as predictions, or load-bearing self-citations are present in the provided text that would reduce the central claims to inputs by construction. The derivation chain for the authoring pipeline and consistency claim remains self-contained against the external benchmarks of the human-subject evaluation.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 1 invented entities

Based on abstract only; the central claim rests on standard assumptions of generative model training and the existence of a perceptually meaningful shared latent space without independent verification beyond the described user study.

axioms (2)
  • domain assumption Generative models (AR and diffusion) can be conditioned on language embeddings to produce outputs aligned with semantic descriptions.
    Invoked implicitly in the system description for turning prompts into textures.
  • domain assumption Participant ratings on real-material anchors provide a valid projection for measuring perceptual consistency in generated textures.
    Used to verify that the latent encodes meaningful structure for roughness, slipperiness, and hardness.
invented entities (1)
  • Shared language-aligned latent space no independent evidence
    purpose: To link haptic and visual generative models so one prompt produces semantically consistent outputs across modalities.
    Postulated as the mechanism enabling coordinated multimodal textures; no independent evidence outside the user study is described.

pith-pipeline@v0.9.0 · 5550 in / 1407 out tokens · 25440 ms · 2026-05-10T18:31:13.844847+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

Reference graph

Works this paper leans on

82 extracted references · 6 canonical work pages · 2 internal anchors

  1. [1]

    Rechap: an interactive recommender system for navigating a large number of mid- air haptic designs,

    K. Theivendran, A. Wu, W. Frier, and O. Schneider, “Rechap: an interactive recommender system for navigating a large number of mid- air haptic designs,”IEEE Transactions on Haptics, vol. 17, no. 2, pp. 165–176, 2023

    2023

  2. [2]

    Genartist: Multimodal llm as an agent for unified image generation and editing,

    Z. Wang, A. Li, Z. Li, and X. Liu, “Genartist: Multimodal llm as an agent for unified image generation and editing,”Advances in Neural Information Processing Systems, vol. 37, pp. 128 374–128 395, 2024

    2024

  3. [3]

    Creating realistic virtual textures from contact acceleration data,

    J. M. Romano and K. J. Kuchenbecker, “Creating realistic virtual textures from contact acceleration data,”IEEE Transactions on Haptics, vol. 5, no. 2, pp. 109–119, 2011

    2011

  4. [4]

    Haptics in minimally invasive surgical simulation and training,

    C. Basdogan, S. De, J. Kim, M. Muniyandi, H. Kim, and M. A. Srinivasan, “Haptics in minimally invasive surgical simulation and training,”IEEE Computer Graphics and Applications, vol. 24, no. 2, pp. 56–64, 2004

    2004

  5. [5]

    A review of virtual reality and haptics for product assembly (part 1): rigid parts,

    P. Xia, A. M. Lopes, and M. T. Restivo, “A review of virtual reality and haptics for product assembly (part 1): rigid parts,”Assembly Automation, vol. 33, no. 1, pp. 68–77, 2013

    2013

  6. [6]

    Vibrotactile display: Perception, technology, and applications,

    S. Choi and K. J. Kuchenbecker, “Vibrotactile display: Perception, technology, and applications,”Proceedings of the IEEE, vol. 101, no. 9, pp. 2093–2104, 2012

    2093

  7. [7]

    Modeling and rendering realistic textures from unconstrained tool-surface interactions,

    H. Culbertson, J. Unwin, and K. J. Kuchenbecker, “Modeling and rendering realistic textures from unconstrained tool-surface interactions,” IEEE Transactions on Haptics, vol. 7, no. 3, pp. 381–393, 2014

    2014

  8. [8]

    Vibviz: Organizing, visualizing and navigating vibration libraries,

    H. Seifi, K. Zhang, and K. E. MacLean, “Vibviz: Organizing, visualizing and navigating vibration libraries,” inIEEE World Haptics Conference (WHC), 2015, pp. 254–259

    2015

  9. [9]

    One hundred data-driven haptic texture models and open-source methods for rendering on 3d objects,

    H. Culbertson, J. J. L. Delgado, and K. J. Kuchenbecker, “One hundred data-driven haptic texture models and open-source methods for rendering on 3d objects,” inIEEE Haptics Symposium (HAPTICS), 2014, pp. 319– 325

    2014

  10. [10]

    Psychophysical dimensions of tactile perception of textures,

    S. Okamoto, H. Nagano, and Y . Yamada, “Psychophysical dimensions of tactile perception of textures,”IEEE Transactions on Haptics, vol. 6, no. 1, pp. 81–93, 2012

    2012

  11. [11]

    Vibrotactile signal generation from texture images or attributes using generative adversarial network,

    Y . Ujitoko and Y . Ban, “Vibrotactile signal generation from texture images or attributes using generative adversarial network,” inInterna- tional conference on human haptic sensing and touch enabled computer applications. Springer, 2018, pp. 25–36

    2018

  12. [12]

    Hapticgen: Generative text-to-vibration model for streamlining haptic design,

    Y . Sung, K. John, S. H. Yoon, and H. Seifi, “Hapticgen: Generative text-to-vibration model for streamlining haptic design,” inProceedings of the CHI Conference on Human Factors in Computing Systems, 2025, pp. 1–24

    2025

  13. [13]

    Tactual perception of material properties,

    W. M. B. Tiest, “Tactual perception of material properties,”Vision research, vol. 50, no. 24, pp. 2775–2782, 2010

    2010

  14. [14]

    Authoring new haptic textures based on interpolation of real textures in affective space,

    W. Hassan, A. Abdulali, and S. Jeon, “Authoring new haptic textures based on interpolation of real textures in affective space,”IEEE Trans- actions on Industrial Electronics, vol. 67, no. 1, pp. 667–676, 2020

    2020

  15. [15]

    Integrating texture models through regression of vibration and texture characteristics,

    K. Tozuka, B. Poitrimol, G. Sasaki, K. Kobayashi, and H. Igarashi, “Integrating texture models through regression of vibration and texture characteristics,”ROBOMECH Journal, vol. 12, no. 1, p. 25, 2025

    2025

  16. [16]

    Preference-driven texture modeling through interactive generation and search,

    S. Lu, M. Zheng, M. C. Fontaine, S. Nikolaidis, and H. Culbertson, “Preference-driven texture modeling through interactive generation and search,”IEEE Transactions on Haptics, vol. 15, no. 3, pp. 508–520, 2022

    2022

  17. [17]

    Learning an action- conditional model for haptic texture generation,

    N. Heravi, W. Yuan, A. M. Okamura, and J. Bohg, “Learning an action- conditional model for haptic texture generation,” inIEEE International Conference on Robotics and Automation (ICRA), 2020, pp. 11 088– 11 095

    2020

  18. [18]

    Learning transferable visual models from natural language supervi- sion,

    A. Radford, J. W. Kim, C. Hallacy, A. Ramesh, G. Goh, S. Agarwal, G. Sastry, A. Askell, P. Mishkin, J. Clark, G. Krueger, and I. Sutskever, “Learning transferable visual models from natural language supervi- sion,” inProceedings of the International Conference on Machine Learning (ICML), 2021

    2021

  19. [19]

    Haptics: The present and future of artificial touch sensation,

    H. Culbertson, S. B. Schorr, and A. M. Okamura, “Haptics: The present and future of artificial touch sensation,”Annual review of control, robotics, and autonomous systems, vol. 1, no. 1, pp. 385–409, 2018

    2018

  20. [20]

    Twenty years of world haptics: Retrospective and future directions,

    J. E. Colgate, L. A. Jones, and H. Z. Tan, “Twenty years of world haptics: Retrospective and future directions,”IEEE Transactions on Haptics, vol. 18, no. 3, pp. 452–455, 2025

    2025

  21. [21]

    Observations on active touch

    J. J. Gibson, “Observations on active touch.”Psychological review, vol. 69, no. 6, p. 477, 1962

    1962

  22. [22]

    Visual and tactual texture perception: Intersensory cooperation,

    M. A. Heller, “Visual and tactual texture perception: Intersensory cooperation,”Perception & psychophysics, vol. 31, no. 4, pp. 339–344, 1982

    1982

  23. [23]

    Merging the senses into a robust percept,

    M. O. Ernst and H. H. B ¨ulthoff, “Merging the senses into a robust percept,”Trends in Cognitive Sciences, vol. 8, no. 4, pp. 162–169, 2004

    2004

  24. [24]

    Multisensory texture perception,

    R. L. Klatzky and S. J. Lederman, “Multisensory texture perception,” in Multisensory object perception in the primate brain. Springer, 2010, pp. 211–230

    2010

  25. [25]

    Can you see what you feel? color and folding properties affect visual–tactile material discrimination of fabrics,

    B. Xiao, W. Bi, X. Jia, H. Wei, and E. H. Adelson, “Can you see what you feel? color and folding properties affect visual–tactile material discrimination of fabrics,”Journal of Vision, vol. 16, no. 3, pp. 34–34, 2016

    2016

  26. [26]

    Survey of tex- ture mapping techniques for representing and rendering volumetric mesostructure,

    C. Koniaris, D. Cosker, X. Yang, and K. Mitchell, “Survey of tex- ture mapping techniques for representing and rendering volumetric mesostructure,”Journal of Computer Graphics Techniques, 2014

    2014

  27. [27]

    Reflectance and texture of real-world surfaces,

    K. J. Dana, B. van Ginneken, S. K. Nayar, and J. J. Koenderink, “Reflectance and texture of real-world surfaces,”ACM Trans. Graph., vol. 18, no. 1, p. 1–34, Jan. 1999

    1999

  28. [28]

    Photoshape: photorealistic materials for large-scale shape collections,

    K. Park, K. Rematas, A. Farhadi, and S. M. Seitz, “Photoshape: photorealistic materials for large-scale shape collections,”ACM Trans. Graph., vol. 37, no. 6, Dec. 2018

    2018

  29. [29]

    Bidirectional visual- tactile cross-modal generation using latent feature space flow model,

    Y . Fang, X. Zhang, W. Xu, G. Liu, and J. Zhao, “Bidirectional visual- tactile cross-modal generation using latent feature space flow model,” Neural Networks, vol. 172, p. 106088, 2024

    2024

  30. [30]

    Natural scenes in tactile texture,

    L. R. Manfredi, H. P. Saal, K. J. Brown, M. C. Zielinski, J. F. Dammann III, V . S. Polashock, and S. J. Bensmaia, “Natural scenes in tactile texture,”Journal of neurophysiology, vol. 111, no. 9, pp. 1792– 1802, 2014

    2014

  31. [31]

    Signal processing for haptic surface modeling: A review,

    A. L. Stefani, N. Bisagno, A. Rosani, N. Conci, and F. De Natale, “Signal processing for haptic surface modeling: A review,”Signal Processing: Image Communication, p. 117338, 2025

    2025

  32. [32]

    A systematic review of haptic texture reproduction technology,

    S. Chen, T. Yuan, L. Xu, W. Ru, and D. Wang, “A systematic review of haptic texture reproduction technology,”Intelligence & Robotics, vol. 5, no. 3, pp. 607–30, 2025

    2025

  33. [33]

    Reality-based models for vibration feedback in virtual environments,

    A. Okamura, M. Cutkosky, and J. Dennerlein, “Reality-based models for vibration feedback in virtual environments,”IEEE/ASME Transactions on Mechatronics, vol. 6, no. 3, pp. 245–252, 2001

    2001

  34. [34]

    Dynamic simulation of tool- mediated texture interaction,

    C. G. McDonald and K. J. Kuchenbecker, “Dynamic simulation of tool- mediated texture interaction,” in2013 World Haptics Conference (WHC), 2013, pp. 307–312

    2013

  35. [35]

    Generating haptic texture models from unconstrained tool-surface in- teractions,

    H. Culbertson, J. Unwin, B. E. Goodman, and K. J. Kuchenbecker, “Generating haptic texture models from unconstrained tool-surface in- teractions,” inWorld Haptics Conference (WHC), 2013, pp. 295–300

    2013

  36. [36]

    Data-driven mod- eling of isotropic haptic textures using frequency-decomposed neural networks,

    S. Shin, R. H. Osgouei, K.-D. Kim, and S. Choi, “Data-driven mod- eling of isotropic haptic textures using frequency-decomposed neural networks,” inIEEE World Haptics Conference (WHC), 2015, pp. 131– 138

    2015

  37. [37]

    Data-driven haptic texture modeling and rendering based on deep spatio-temporal networks,

    J. B. Joolee and S. Jeon, “Data-driven haptic texture modeling and rendering based on deep spatio-temporal networks,”IEEE Transactions on Haptics, vol. 15, no. 1, pp. 62–67, 2022

    2022

  38. [38]

    Development and evaluation of a learning-based model for real-time haptic texture rendering,

    N. Heravi, H. Culbertson, A. M. Okamura, and J. Bohg, “Development and evaluation of a learning-based model for real-time haptic texture rendering,”IEEE Transactions on Haptics, vol. 17, no. 4, pp. 705–716, 2024

    2024

  39. [39]

    Time series diffusion method: A denoising diffusion probabilistic model for vibration signal generation,

    X. Yi, J. Wang, and P. Sun, “Time series diffusion method: A denoising diffusion probabilistic model for vibration signal generation,”arXiv preprint arXiv:2303.01234, 2023

    work page arXiv 2023

  40. [40]

    Improving contact realism through event-based haptic feedback,

    K. J. Kuchenbecker, J. Fiene, and G. Niemeyer, “Improving contact realism through event-based haptic feedback,”IEEE transactions on visualization and computer graphics, vol. 12, no. 2, pp. 219–230, 2006

    2006

  41. [41]

    T-pad: Tactile pattern display through variable friction reduction,

    L. Winfield, J. Glassmire, J. E. Colgate, and M. Peshkin, “T-pad: Tactile pattern display through variable friction reduction,” inSecond Joint EuroHaptics Conference and Symposium on Haptic Interfaces for Virtual Environment and Teleoperator Systems (WHC’07), 2007, pp. 421–426

    2007

  42. [42]

    Fingertip friction modulation due to electrostatic attraction,

    D. J. Meyer, M. A. Peshkin, and J. E. Colgate, “Fingertip friction modulation due to electrostatic attraction,” in2013 World Haptics Conference (WHC), 2013, pp. 43–48

    2013

  43. [43]

    The slip hypothesis: tactile perception and its neuronal bases,

    C. Schwarz, “The slip hypothesis: tactile perception and its neuronal bases,”Trends in neurosciences, vol. 39, no. 7, pp. 449–462, 2016

    2016

  44. [44]

    Towards multisensory perception: Modeling and rendering sounds of tool-surface interactions,

    S. Lu, Y . Chen, and H. Culbertson, “Towards multisensory perception: Modeling and rendering sounds of tool-surface interactions,”IEEE Transactions on Haptics, vol. 13, no. 1, pp. 94–101, 2020

    2020

  45. [45]

    Auditory materiality: Exploring auditory effects on material perception in virtual reality,

    B. Zhong, S. Je, and M. Z. Sagesser, “Auditory materiality: Exploring auditory effects on material perception in virtual reality,” in2025 IEEE Conference on Virtual Reality and 3D User Interfaces Abstracts and Workshops (VRW), 2025, pp. 659–662

    2025

  46. [46]

    Thermalbitdisplay: Haptic display providing thermal feedback perceived differently depend- ing on body parts,

    A. Niijima, T. Takeda, T. Mukouchi, and T. Satou, “Thermalbitdisplay: Haptic display providing thermal feedback perceived differently depend- ing on body parts,” inExtended Abstracts of the CHI Conference on Human Factors in Computing Systems, 2020, pp. 1–8

    2020

  47. [47]

    Three-axis pneumatic haptic display for the mechanical and thermal stimulation of a human finger pad,

    E.-H. Lee, S.-H. Kim, and K.-S. Yun, “Three-axis pneumatic haptic display for the mechanical and thermal stimulation of a human finger pad,” inActuators, vol. 10, no. 3. MDPI, 2021, p. 60

    2021

  48. [48]

    Fluid reality: High-resolution, untethered haptic gloves using electroosmotic pump arrays,

    V . Shen, T. Rae-Grant, J. Mullenbach, C. Harrison, and C. Shultz, “Fluid reality: High-resolution, untethered haptic gloves using electroosmotic pump arrays,” inProceedings of the 36th Annual ACM Symposium on User Interface Software and Technology, 2023, pp. 1–20

    2023

  49. [49]

    Shape-it: Exploring text-to-shape-display for generative shape-changing behaviors with llms,

    W. Qian, C. Gao, A. Sathya, R. Suzuki, and K. Nakagaki, “Shape-it: Exploring text-to-shape-display for generative shape-changing behaviors with llms,” inProceedings of the 37th Annual ACM Symposium on User Interface Software and Technology, 2024, pp. 1–29

    2024

  50. [50]

    Texturegan: Controlling deep image synthesis with texture patches,

    W. Xian, P. Sangkloy, V . Agrawal, A. Raj, J. Lu, C. Fang, F. Yu, and J. Hays, “Texturegan: Controlling deep image synthesis with texture patches,” inProceedings of the IEEE conference on computer vision and pattern recognition, 2018, pp. 8456–8465

    2018

  51. [51]

    Diffusion models: A comprehensive survey of methods and applications,

    L. Yang, Z. Zhang, Y . Song, S. Hong, R. Xu, Y . Zhao, W. Zhang, B. Cui, and M.-H. Yang, “Diffusion models: A comprehensive survey of methods and applications,”ACM computing surveys, vol. 56, no. 4, pp. 1–39, 2023

    2023

  52. [52]

    Sequential gallery for interactive visual design optimization,

    Y . Koyama, I. Sato, and M. Goto, “Sequential gallery for interactive visual design optimization,”ACM Transactions on Graphics (TOG), vol. 39, no. 4, pp. 88–1, 2020

    2020

  53. [53]

    Text2tex: Text-driven texture synthesis via diffusion models,

    D. Z. Chen, Y . Siddiqui, H.-Y . Lee, S. Tulyakov, and M. Nießner, “Text2tex: Text-driven texture synthesis via diffusion models,” inPro- ceedings of the IEEE/CVF international conference on computer vision, 2023, pp. 18 558–18 568

    2023

  54. [54]

    Vibrotactile signal generation from texture images or attributes using generative adversarial network,

    Y . Ujitoko and Y . Ban, “Vibrotactile signal generation from texture images or attributes using generative adversarial network,” inHaptics: Science, Technology, and Applications, D. Prattichizzo, H. Shinoda, H. Z. Tan, E. Ruffaldi, and A. Frisoli, Eds. Cham: Springer International Publishing, 2018, pp. 25–36

    2018

  55. [55]

    TEXasGAN: Tactile texture exploration and synthesis system using generative adversarial network,

    M. Zhang, S. Terui, Y . Makino, and H. Shinoda, “TEXasGAN: Tactile texture exploration and synthesis system using generative adversarial network,”arXiv preprint arXiv:2407.11467, 2024

    work page arXiv 2024

  56. [56]

    Tactstyle: Generating tactile textures with generative ai for digital fabrication,

    F. Faruqi, M. Perroni-Scharf, J. S. Walia, Y . Zhu, S. Feng, D. Degraen, and S. Mueller, “Tactstyle: Generating tactile textures with generative ai for digital fabrication,” inProceedings of the 2025 CHI Conference on Human Factors in Computing Systems, 2025, pp. 1–16

    2025

  57. [57]

    Gan-based image-to-friction generation for tactile simulation of fabric material,

    S. Cai, L. Zhao, Y . Ban, T. Narumi, Y . Liu, and K. Zhu, “Gan-based image-to-friction generation for tactile simulation of fabric material,” Computers & Graphics, vol. 104, pp. 219–228, 2022

    2022

  58. [58]

    Vis2hap: Vision-based haptic rendering by cross-modal generation,

    G. Cao, J. Jiang, N. Mao, D. Bollegala, M. Li, and S. Luo, “Vis2hap: Vision-based haptic rendering by cross-modal generation,” inIEEE International Conference on Robotics and Automation (ICRA), 2023, pp. 12 443–12 449

    2023

  59. [59]

    Cm-avae: Cross-modal adversarial variational autoencoder for visual-to-tactile data generation,

    Q. Xi, F. Wang, L. Tao, H. Zhang, X. Jiang, and J. Wu, “Cm-avae: Cross-modal adversarial variational autoencoder for visual-to-tactile data generation,”IEEE Robotics and Automation Letters, vol. 9, no. 6, pp. 5214–5221, 2024

    2024

  60. [60]

    Method for audio-to-tactile cross- modality generation based on residual u-net,

    H. Zhan, J. Chen, and L. Huang, “Method for audio-to-tactile cross- modality generation based on residual u-net,”IEEE Transactions on Instrumentation and Measurement, vol. 73, pp. 1–10, 2024

    2024

  61. [61]

    Cross-modal generation of tactile friction coefficient from audio and visual measurements by transformer,

    R. Song, X. Sun, and G. Liu, “Cross-modal generation of tactile friction coefficient from audio and visual measurements by transformer,”IEEE Transactions on Instrumentation and Measurement, vol. 72, pp. 1–11, 2023

    2023

  62. [62]

    High- resolution image synthesis with latent diffusion models,

    R. Rombach, A. Blattmann, D. Lorenz, P. Esser, and B. Ommer, “High- resolution image synthesis with latent diffusion models,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recogni- tion (CVPR), 2022, pp. 10 684–10 695

    2022

  63. [63]

    Hierarchical Text-Conditional Image Generation with CLIP Latents

    A. Ramesh, P. Dhariwal, A. Nichol, C. Chu, and M. Chen, “Hierarchical text-conditional image generation with clip latents,”arXiv preprint arXiv:2204.06125, 2022

    work page internal anchor Pith review Pith/arXiv arXiv 2022

  64. [64]

    Photorealistic text-to-image diffusion models with deep language understanding,

    C. Saharia, W. Chan, S. Saxena, L. Li, J. Whang, E. L. Denton, K. Ghasemipour, R. Gontijo Lopes, B. Karagol Ayan, T. Salimans et al., “Photorealistic text-to-image diffusion models with deep language understanding,”Advances in neural information processing systems, vol. 35, pp. 36 479–36 494, 2022

    2022

  65. [65]

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

    C.-H. Lin, J. Gao, L. Tang, T. Takikawa, X. Zeng, X. Huang, K. Kreis, S. Fidler, M.-Y . Liu, and T.-Y . Lin, “Magic3d: High-resolution text-to- 3d content creation,” inProceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2023, pp. 300–309

    2023

  66. [66]

    Dream3d: Zero-shot text-to-3d synthesis using 3d shape prior and text- to-image diffusion models,

    J. Xu, X. Wang, W. Cheng, Y .-P. Cao, Y . Shan, X. Qie, and S. Gao, “Dream3d: Zero-shot text-to-3d synthesis using 3d shape prior and text- to-image diffusion models,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2023, pp. 20 908–20 918

    2023

  67. [67]

    Audiogen: Textually guided audio generation,

    F. Kreuk, G. Synnaeve, A. Polyak, U. Singer, A. D ´efossez, J. Copet, D. Parikh, Y . Taigman, and Y . Adi, “Audiogen: Textually guided audio generation,”arXiv preprint arXiv:2209.15352, 2022

    work page arXiv 2022

  68. [68]

    MusicLM: Generating Music From Text

    A. Agostinelli, T. I. Denk, Z. Borsos, J. Engel, M. Verzetti, A. Caillon, Q. Huang, A. Jansen, A. Roberts, M. Tagliasacchiet al., “Musiclm: Generating music from text,”arXiv preprint arXiv:2301.11325, 2023

    work page internal anchor Pith review arXiv 2023

  69. [69]

    Audioldm 2: Learning holistic audio generation with self-supervised pretraining,

    H. Liu, Y . Yuan, X. Liu, X. Mei, Q. Kong, Q. Tian, Y . Wang, W. Wang, Y . Wang, and M. D. Plumbley, “Audioldm 2: Learning holistic audio generation with self-supervised pretraining,”IEEE/ACM Transactions on Audio, Speech, and Language Processing, vol. 32, pp. 2871–2883, 2024

    2024

  70. [70]

    Texttoucher: Fine-grained text-to-touch generation,

    J. Tu, H. Fu, F. Yang, H. Zhao, C. Zhang, and H. Qian, “Texttoucher: Fine-grained text-to-touch generation,” inProceedings of the AAAI Conference on Artificial Intelligence, vol. 39, 2025, pp. 7455–7463

    2025

  71. [71]

    Text-driven generative framework for multimodal visual and haptic texture synthesis,

    M. Naeem, M. I. Awan, and S. Jeon, “Text-driven generative framework for multimodal visual and haptic texture synthesis,” in2025 IEEE World Haptics Conference (WHC), 2025, pp. 140–146

    2025

  72. [72]

    Penn haptic texture toolkit: A collection of textures for haptic rendering,

    H. Culbertson and K. J. Kuchenbecker, “Penn haptic texture toolkit: A collection of textures for haptic rendering,” inIEEE Haptics Symposium (HAPTICS), 2014, pp. 99–106

    2014

  73. [73]

    Importance of matching physical friction, hardness, and texture in creating realistic haptic virtual surfaces,

    ——, “Importance of matching physical friction, hardness, and texture in creating realistic haptic virtual surfaces,”IEEE Transactions on Haptics, vol. 10, no. 1, pp. 63–74, 2016

    2016

  74. [74]

    Chatgpt (gpt-5),

    OpenAI, “Chatgpt (gpt-5),” https://chat.openai.com, 2025, large lan- guage model developed by OpenAI. URL: https://chat.openai.com

    2025

  75. [75]

    beta-vae: Learning basic visual concepts with a constrained variational framework,

    I. Higgins, L. Matthey, A. Pal, C. Burgess, X. Glorot, M. Botvinick, S. Mohamed, and A. Lerchner, “beta-vae: Learning basic visual concepts with a constrained variational framework,” inInternational conference on learning representations, 2017

    2017

  76. [76]

    Stable diffusion v2-base — stabilityai / hugging face,

    “Stable diffusion v2-base — stabilityai / hugging face,” https://huggingface.co/stabilityai/stable-diffusion-2-base, 2025, accessed: 2025-09-30

    2025

  77. [77]

    Textures — poly haven,

    “Textures — poly haven,” https://polyhaven.com/textures, 2025, ac- cessed: 2025-09-30

    2025

  78. [78]

    Clustergan: Latent space clustering in generative adversarial networks,

    S. Mukherjee, H. Asnani, E. Lin, and S. Kannan, “Clustergan: Latent space clustering in generative adversarial networks,” inProceedings of the AAAI conference on artificial intelligence, vol. 33, no. 01, 2019, pp. 4610–4617

    2019

  79. [79]

    Unsupervised deep embedding for clustering analysis,

    J. Xie, R. Girshick, and A. Farhadi, “Unsupervised deep embedding for clustering analysis,” inInternational conference on machine learning. PMLR, 2016, pp. 478–487

    2016

  80. [80]

    Gaussian mixture generative adver- sarial networks for diverse datasets, and the unsupervised clustering of images,

    M. Ben-Yosef and D. Weinshall, “Gaussian mixture generative adver- sarial networks for diverse datasets, and the unsupervised clustering of images,”arXiv preprint arXiv:1808.10356, 2018

    work page arXiv 2018

Showing first 80 references.