Pith. sign in

REVIEW 4 major objections 4 minor 1 cited by

VQ-Touch generates high-fidelity tactile images for new sensors using only about 50 real images per sensor, by sharing a learned codebook across sensor families.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

T0 review · deepseek-v4-flash

2026-08-02 01:11 UTC pith:AU56S2NP

load-bearing objection Reasonable engineering, but the cross-sensor story is unvalidated: 'zero-shot' is actually few-shot, and the MAE comparison is confounded. the 4 major comments →

arxiv 2607.14728 v1 pith:AU56S2NP submitted 2026-07-16 cs.CV cs.RO

VQ-Touch: A Data-Efficient Tactile Generation Framework Across Sensors and Scenarios

Kailin Lyu , Long Xiao , Jianing Zeng , Di Wu , Lin Shu , Jie Hao This is my paper
classification cs.CV cs.RO
keywords tactile image generationvision-based tactile sensorsVQGANdiscrete diffusionfew-shot learningcross-sensor generalizationmultimodal generationrepresentation learning
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

VQ-Touch asks whether tactile image generation can be made data-efficient enough to work across different tactile sensors with minimal new data. The paper argues yes: a VQGAN variant adapted to the low-variance, deformation-heavy structure of vision-based tactile images learns a compact codebook; sensors that share design are clustered into families; and a discrete diffusion model generates tactile images conditioned on visual images, other tactile images, or labels. With family data plus very few target images, the framework reconstructs and generates tactile images better than baselines, including a masked-autoencoder baseline trained on 100 times more data. If the claim holds, synthetic touch data becomes practical for training robotic perception where real tactile data is expensive or sensor-specific.

Core claim

The paper's central claim is that a single framework, VQ-Touch, can generate high-fidelity tactile images across sensors and scenarios while needing very little data. The supporting discovery is that representing tactile images with a deformable multi-scale VQGAN codebook, then aligning sensors into families by unsupervised feature clustering, allows a codebook trained on family data to be transferred to a new sensor with only around 50 real images from that sensor mixed at a 1:5 ratio with family data. A discrete diffusion decoder then converts visual images, other tactile images, or semantic labels into tactile images. Across four public datasets spanning GelSight, GelSight Mini, and DIGIT

What carries the argument

DM-VQGAN, a VQGAN encoder-decoder that uses deformable convolutions to learn spatial offsets for irregular deformation patterns and parallel dilated convolutions (rates 1, 2, 4) to fuse multi-scale features, producing discrete codebook representations tailored to tactile images. Sensor-family clustering assigns unseen sensors to a family by computing feature prototypes and nearest-cluster distance, after which few-shot mixed training combines a small number of target-sensor images with family images at a 1:5 ratio to adapt the codebook. A discrete diffusion decoder, operating in the discrete latent space and conditioned through a frozen CLIP encoder for visual inputs, the learned codebook fo

Load-bearing premise

The load-bearing premise is that sensors clustered into the same family share a tactile image representation, so a codebook trained on family data plus only 50 images of a new sensor transfers to that sensor; if family similarity is weak, the reported gains could simply come from the 50 target images.

What would settle it

Re-run the cross-sensor reconstruction with the same 50 images but replace the family training data with equal amounts of unrelated sensor data, or with no family data; if FID/SSIM do not worsen, sensor-family clustering is not the cause of the transfer. Alternatively, test zero-shot reconstruction with 0 target images; failure would contradict the zero-shot label.

Watch this falsifier — get emailed when new claim-graph text bears on it.

If this is right

  • Tactile data collection for a new sensor could shrink from large multi-dataset pretraining to under a hundred real images plus family data, making synthetic data viable for robotics.
  • Vision-to-touch and label-to-touch generation allow tactile training data to be produced from scene images or semantic annotations when a physical sensor is unavailable.
  • Classifiers trained on VQ-Touch-generated images transfer to real recognition tasks with higher accuracy than prior generated images, so the synthetic data carries usable contact information.
  • The same codebook infrastructure supports multiple mainstream vision-based tactile sensor variants, suggesting one model can serve a fleet of sensors rather than per-sensor models.
  • Few-shot mixed training may reduce redundant data collection across overlapping sensor designs.

Where Pith is reading between the lines

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

  • The 'zero-shot transfer' label is optimistic: the reported cross-sensor experiments train on 50 images from the unseen sensor, so the contribution of the family codebook is not isolated; true zero-shot transfer remains untested.
  • If sensor-family clustering really is doing the work, its quality is measurable—clustering accuracy against known sensor lineage, or ablating with random family assignment, would quantify how much sharing matters. The paper does not provide this validation.
  • The conditioning interface is generic enough to extend beyond the three modalities tested; one could condition on force readings or audio, a step the paper does not take.
  • Data efficiency at 100 samples suggests the bottleneck is codebook coverage, not model capacity; a testable extension is scaling latent codebook size against sample count.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 4 minor

Summary. The paper proposes VQ-Touch, a tactile image generation framework combining a VQGAN-based encoder (DM-VQGAN) with deformable and dilated convolutions, a sensor-family clustering scheme, few-shot mixed training, and a discrete diffusion decoder for multimodal conditioning. The stated contributions are (i) DM-VQGAN as an efficient tactile representation learner, (ii) a unified multimodal conditional generation framework, (iii) sensor-family-level modeling for cross-sensor transfer, and (iv) state-of-the-art reconstruction and generation performance under limited data. Experiments are reported on four tactile datasets with comparisons to MAE, VQGAN, UniTouch, PixArt-α, and TextToucher.

Significance. If the claims were fully supported, VQ-Touch would be a practically valuable contribution to tactile sensing, where data collection is expensive and sensor diversity is high. The idea of sharing representations across sensor families via clustering and few-shot adaptation is interesting and potentially impactful. The paper also includes a useful unified conditioning interface for vision, tactile, and label inputs. However, the current evidence is not sufficient to establish these claims: the main reconstruction comparison is confounded, the cross-sensor generalization experiment is qualitative and trains on target-sensor data, and the sensor-family clustering is not validated at all. The architectural ingredients are largely borrowed from prior VQGAN and discrete diffusion work, so the novelty rests on the cross-sensor claims, which are currently unsupported.

major comments (4)
  1. [Table I, Sec. III-B] The comparison between MAE and VQGAN/DM-VQGAN is not apples-to-apples. MAE is trained for masked-input reconstruction with mask ratios 0.25, 0.5, and 0.75, while VQGAN and DM-VQGAN are trained for full-image autoencoding. The conclusion that 'MAE fails' and that DM-VQGAN with 100 samples outperforms MAE with 10k samples does not follow from this table, because the two methods are optimizing different tasks. To support the claim, MAE should also be evaluated on full-image reconstruction (or otherwise matched to the same training objective), or the comparison should be reframed.
  2. [Sec. III-B, Fig. 5] The cross-sensor generalization claim is not quantitatively supported. The text and caption state that 50 images from unseen sensors are combined with family data at a 1:5 ratio to train DM-VQGAN. This is a few-shot setup, not zero-shot, yet the abstract and introduction claim 'zero-shot transfer.' More importantly, Fig. 5 shows only qualitative reconstructions with no FID/SSIM/LPIPS numbers and no baseline. There is no ablation comparing (a) training only on the 50 target-sensor images, (b) training only on family data, and (c) the proposed mixed training. Without such an ablation, the contribution of the sensor-family representation to cross-sensor performance is not established.
  3. [Sec. II-B, Eq. (5)] The sensor-family clustering is a central novelty of the paper, but it is never validated experimentally. No cluster assignments are reported, no cluster validity metric is given, and there is no sensitivity analysis with respect to the feature extractor, the clustering algorithm, or the number of clusters. The claim that sensors with 'shared intrinsic properties' form meaningful families is a hypothesis; the paper provides no evidence that the proposed clustering produces families that improve transfer over random grouping or over training directly on the new sensor's data.
  4. [Tables II and IV] The claim that VQ-Touch 'outperforms state-of-the-art models' is weakened by incomplete experimental description. The CTTP metric in Table II is not defined, and no variance or significance measures are reported even though the text states experiments were repeated three times. Table IV compares only against UniTouch, omitting the stronger recent baselines that appear in Table II. Without a clear protocol and a consistent set of baselines across all tasks, the SOTA claim is not fully substantiated.
minor comments (4)
  1. [Throughout] Typos and grammatical issues: 'm ulti-scale' in Sec. II-A, 'unimodal' appears in the conclusion where 'multimodal' seems intended, and several sentences are run-on. A careful proofread is needed.
  2. [Sec. II-C] The discrete diffusion decoder is described at a high level, but no training objective, diffusion schedule, or algorithm for the discrete diffusion process is provided. The paper cites prior work, but since the unified conditioning interface is a claimed contribution, more technical detail is needed for reproducibility.
  3. [Fig. 2] Figure 2 is dense and hard to parse, especially the 'Clustering' and 'Mixed Training' components. The relationship between sensor-specific codebooks and the shared codebook is unclear from the figure alone.
  4. [Sec. III-A] The datasets are listed but no statistics (image counts per sensor, resolution, train/test splits) are given. This makes it difficult to assess the data-efficiency claims.

Circularity Check

0 steps flagged

No significant circularity: the reconstruction/generation pipeline is self-contained; the cross-sensor 'unseen' protocol is an evidence limitation, not a circular reduction.

full rationale

The paper's derivation is not circular. DM-VQGAN minimizes standard external losses (Eq. 2: reconstruction, VQ commitment, perceptual, adversarial), with the target being faithful reconstruction of tactile images; no parameter is fitted to the evaluation metrics. The discrete diffusion decoder follows standard VQ-diffusion (Eqs. 6-7), and the sensor-family clustering is a heuristic grouping with a defined distance (Eq. 5) that is not defined in terms of the reconstruction target. Self-citations [15-17,21] appear only as related-work pointers (texture recognition, tactile data scarcity) and are not load-bearing for the framework's derivation or for the claimed SOTA numbers. The main weakness is in Sec. III-B/Fig. 5: 'a small number of tactile images from unseen sensors are combined with existing datasets from their corresponding sensor families to train DM-VQGAN' and the evidence is qualitative. This means the cross-sensor demonstration is actually few-shot adaptation, not zero-shot transfer, and the family-data contribution is not isolated by an ablation. This is an evidentiary/reporting problem about generalization strength, not a case where the prediction is identical to the input by construction; therefore the circularity score remains low.

Axiom & Free-Parameter Ledger

4 free parameters · 5 axioms · 0 invented entities

The framework rests on standard generative-modeling axioms and a small number of hand-chosen hyperparameters. The most load-bearing, unvalidated assumption is the sensor-family hypothesis, which is ad hoc and essential to the cross-sensor generalization claim. No new physical entities or conserved quantities are introduced.

free parameters (4)
  • Few-shot mix ratio = 1:5 (target:family)
    Chosen by hand for cross-sensor training; no sensitivity analysis or justification.
  • Target-sensor sample count = 50 images
    Used in the cross-sensor experiment; the presence of 50 target images in training undermines the zero-shot claim.
  • Dilation rates = {1,2,4}
    Architecture choice for the multi-scale fusion module; no ablation is provided.
  • Commitment weight beta = 0.25
    Borrowed from prior VQGAN work; not tuned or ablated for tactile data.
axioms (5)
  • domain assumption Vision-based tactile images have low intrinsic variance and discard color, making a discrete VQGAN latent space suitable (Sec. II-A).
    Motivates the codebook approach; if false, the discrete representation may not capture important tactile structure.
  • ad hoc to paper Sensors with shared intrinsic properties but different environmental conditions form a family whose representations can be jointly learned (Sec. II-B).
    Core to the cross-sensor generalization claim; no quantitative evidence validates this clustering hypothesis.
  • ad hoc to paper Clustering sensor features via Eq. 5 yields meaningful families for unseen sensors.
    The clustering is not evaluated; the assignment of an unseen sensor to a family is assumed to be correct.
  • domain assumption Reconstruction quality (FID/SSIM) and classifier accuracy on generated images are good proxies for usefulness in real tactile tasks.
    The paper uses these metrics to support practical value, but does not validate on a real robot or downstream manipulation task.
  • standard math Standard VQGAN losses (perceptual, adversarial, commitment) transfer appropriately to tactile images.
    Borrowed from image synthesis; the particularities of tactile deformation may require different perceptual weights, which are not investigated.

pith-pipeline@v1.3.0-alltime-deepseek · 7473 in / 9237 out tokens · 81643 ms · 2026-08-02T01:11:05.102864+00:00 · methodology

0 comments
read the original abstract

Tactile image generation significantly reduces the dependency on expensive and wear-prone sensors by synthesizing high-fidelity tactile data, offering an efficient solution for tactile information acquisition in robotic perception and human-machine interaction systems. However, existing methods depend on large-scale, diverse datasets from specific sensors and lack efficient data utilization and robust generalization capabilities, struggling in vision-limited environments. To address this, we introduce VQ-Touch, a tactile generation framework that supports both cross-sensor and multi-scenario applications. Specifically, to efficiently extract complex deformation and texture features from the data, we propose DM-VQGAN, an effective tactile representation learner. Furthermore, we introduce a discrete diffusion decoder with a unified conditioning interface, supporting multimodal generation tasks such as images and labels, and enhances the model's generalization capability through few-shot mixed training, thus achieving compatibility with current mainstream sensors and their variants. Experiments show that VQ-Touch surpasses state-of-the-art methods in multiple tasks.

Figures

Figures reproduced from arXiv: 2607.14728 by Di Wu, Jianing Zeng, Jie Hao, Kailin Lyu, Lin Shu, Long Xiao.

Figure 1
Figure 1. Figure 1: The illustration of tactile generation tasks and comparison of different methods. Leveraging data-efficient tactile representations and a unified discrete diffusion model, VQ-Touch supports few-shot, zero-shot transfer, and cross￾sensor tactile generation. tactile representation that facilitates tactile generation across sensors and scenarios? To address this challenge, we introduce VQ-Touch, a tactile gen… view at source ↗
Figure 2
Figure 2. Figure 2: Overview of the proposed VQ-Touch framework. The framework consists of three stages: (1) learning efficient tactile features using the DM-VQGAN encoder-decoder architecture, which is based on discrete codebooks for representation, (2) acquiring sensor-family-specific codebooks through clustering and few-shot mixed training method, and (3) by leveraging the discrete latent space and encoder learned by DM-VQ… view at source ↗
Figure 3
Figure 3. Figure 3: Latent space visualization and reconstruction com￾parison. (a) shows RGB latent features from three encoders. (b) presents the reconstruction results of MAE with a ViT-Base backbone. (c) shows zero-shot reconstruction by VQ-Touch, demonstrating generalization to unseen objects. TABLE II: Quantitative Results on Tactile Generation Task. Our method achieves the best results on most metrics. Method HCT SSVTP … view at source ↗
Figure 4
Figure 4. Figure 4: and Tab. II, VQ-Touch produces tactile images that maintain strong cross-modal consistency, structural fidelity, and semantic alignment. Finally, classifiers trained on gener￾ated tactile images achieve higher object recognition accuracy than baseline methods (Tab. III), demonstrating the practical value of the generated data for downstream applications. Validation of Transferability and Generalizability a… view at source ↗
Figure 5
Figure 5. Figure 5 [PITH_FULL_IMAGE:figures/full_fig_p004_5.png] view at source ↗

discussion (0)

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

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. FELT: Generating Tactile Signals from Vision for Visuo-Tactile Manipulation

    cs.RO 2026-07 conditional novelty 6.0

    FELT predicts finger pressure maps from RGB images and uses them or their learned features to improve manipulation policies without real tactile sensors at deployment.

Reference graph

Works this paper leans on

27 extracted references · 6 linked inside Pith · cited by 1 Pith paper

  1. [1]

    Exploring modern sensor in robotics: A review

    Vikas Raviraj Acharya and Vidya S Rao. Exploring modern sensor in robotics: A review. In2024 Asia Pacific Conference on Innovation in Technology (APCIT), pages 1–5. IEEE, 2024

    2024

  2. [2]

    More than a feeling: Learning to grasp and regrasp using vision and touch

    Roberto Calandra, Andrew Owens, Dinesh Jayaraman, Justin Lin, Wenzhen Yuan, Jitendra Malik, Edward H Adelson, and Sergey Levine. More than a feeling: Learning to grasp and regrasp using vision and touch. IEEE Robotics and Automation Letters, 3(4):3300–3307, 2018

    2018

  3. [3]

    Multimodal perception for dexterous manipulation

    Guanqun Cao and Shan Luo. Multimodal perception for dexterous manipulation. InTactile Sensing, Skill Learning, and Robotic Dexterous Manipulation, pages 45–58. Elsevier, 2022

    2022

  4. [4]

    Multimodal zero-shot learning for tactile texture recognition.Robotics and Autonomous Systems, 176:104688, 2024

    Guanqun Cao, Jiaqi Jiang, Danushka Bollegala, Min Li, and Shan Luo. Multimodal zero-shot learning for tactile texture recognition.Robotics and Autonomous Systems, 176:104688, 2024

    2024

  5. [5]

    Efficient-vqgan: Towards high-resolution image generation with efficient vision transformers

    Shiyue Cao, Yueqin Yin, Lianghua Huang, Yu Liu, Xin Zhao, Deli Zhao, and Kaigi Huang. Efficient-vqgan: Towards high-resolution image generation with efficient vision transformers. InProceedings of the IEEE/CVF International Conference on Computer Vision, pages 7368–7377, 2023

    2023

  6. [6]

    Deformable convolu- tional networks

    Jifeng Dai, Haozhi Qi, Yuwen Xiong, Yi Li, Guodong Zhang, Han Hu, and Yichen Wei. Deformable convolu- tional networks. InProceedings of the IEEE international conference on computer vision, pages 764–773, 2017

    2017

  7. [7]

    Tam- ing transformers for high-resolution image synthesis

    Patrick Esser, Robin Rombach, and Bjorn Ommer. Tam- ing transformers for high-resolution image synthesis. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 12873–12883, 2021

    2021

  8. [8]

    A touch, vision, and language dataset for multimodal alignment.arXiv preprint arXiv:2402.13232, 2024

    Letian Fu, Gaurav Datta, Huang Huang, William Chung- Ho Panitch, Jaimyn Drake, Joseph Ortiz, Mustafa Mukadam, Mike Lambeta, Roberto Calandra, and Ken Goldberg. A touch, vision, and language dataset for multimodal alignment.arXiv preprint arXiv:2402.13232, 2024

    Pith/arXiv arXiv 2024

  9. [9]

    Generative adversarial networks

    Ian Goodfellow, Jean Pouget-Abadie, Mehdi Mirza, Bing Xu, David Warde-Farley, Sherjil Ozair, Aaron Courville, and Yoshua Bengio. Generative adversarial networks. Communications of the ACM, 63(11):139–144, 2020

    2020

  10. [10]

    Vector quantized diffusion model for text-to-image syn- thesis

    Shuyang Gu, Dong Chen, Jianmin Bao, Fang Wen, Bo Zhang, Dongdong Chen, Lu Yuan, and Baining Guo. Vector quantized diffusion model for text-to-image syn- thesis. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 10696– 10706, 2022

    2022

  11. [11]

    Masked autoencoders are scalable vision learners

    Kaiming He, Xinlei Chen, Saining Xie, Yanghao Li, Piotr Doll´ar, and Ross Girshick. Masked autoencoders are scalable vision learners. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 16000–16009, 2022

    2022

  12. [12]

    Sensing and recognizing surface textures using a gelsight sensor

    Rui Li and Edward H Adelson. Sensing and recognizing surface textures using a gelsight sensor. InProceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pages 1241–1247, 2013

    2013

  13. [13]

    Gelsight fin ray: Incorporating tactile sensing into a soft compliant robotic gripper

    Sandra Q Liu and Edward H Adelson. Gelsight fin ray: Incorporating tactile sensing into a soft compliant robotic gripper. In2022 IEEE 5th International Conference on Soft Robotics (RoboSoft), pages 925–931. IEEE, 2022

    2022

  14. [14]

    M2curl: Sample-efficient multimodal reinforcement learning via self-supervised representation learning for robotic manipulation

    Fotios Lygerakis, Vedant Dave, and Elmar Rueck- ert. M2curl: Sample-efficient multimodal reinforcement learning via self-supervised representation learning for robotic manipulation. In2024 21st International Confer- ence on Ubiquitous Robots (UR), pages 490–497. IEEE, 2024

    2024

  15. [15]

    Tacreasoner: A dynamic tactile-language frame- work for interactive reasoning in real-world scenarios

    Kailin Lyu, Di Wu, Long Xiao, Jianning Zeng, Jianwei He, Chang Lin, Lianyu Hu, Lin Shu, Jie Hao, and Ce Hao. Tacreasoner: A dynamic tactile-language frame- work for interactive reasoning in real-world scenarios. arXiv preprint arXiv:2607.05131, 2026

    Pith/arXiv arXiv 2026

  16. [16]

    Touchthinker: Scaling tactile com- monsense reasoning to the open world with large-scale data and action-aware representation.arXiv preprint arXiv:2606.11637, 2026

    Kailin Lyu, Di Wu, Pengwei Zhang, Yuhang Zheng, Yingxin Lai, Long Xiao, Kangyi Wu, Pengna Li, Chen Gao, Lianyu Hu, et al. Touchthinker: Scaling tactile com- monsense reasoning to the open world with large-scale data and action-aware representation.arXiv preprint arXiv:2606.11637, 2026

    Pith/arXiv arXiv 2026

  17. [17]

    Touchformer: A robust transformer-based framework for multimodal material perception

    Kailin Lyu, Long Xiao, Jianing Zeng, Junhao Dong, Xuexin Liu, Zhuojun Zou, Haoyue Yang, Lin Shu, and Jie Hao. Touchformer: A robust transformer-based framework for multimodal material perception. InPro- ceedings of the AAAI Conference on Artificial Intelli- gence, volume 40, pages 18496–18504, 2026

    2026

  18. [18]

    Learning transferable visual models from natural lan- guage supervision

    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 lan- guage supervision. InInternational conference on ma- chine learning, pages 8748–8763. PmLR, 2021

    2021

  19. [19]

    Tactile data generation and applications based on visuo-tactile sensors: A review

    Yuhao Sun, Ning Cheng, Shixin Zhang, Wenzhuang Li, Lingyue Yang, Shaowei Cui, Huaping Liu, Fuchun Sun, Jianwei Zhang, Guo Di, et al. Tactile data generation and applications based on visuo-tactile sensors: A review. Information Fusion, page 103162, 2025

    2025

  20. [20]

    Visuotactile affordances for cloth manipulation with local control

    Neha Sunil, Shaoxiong Wang, Yu She, Edward Adelson, and Alberto Rodriguez Garcia. Visuotactile affordances for cloth manipulation with local control. InConference on Robot Learning, pages 1596–1606. PMLR, 2023

    2023

  21. [21]

    Tacexpert: A pseudo-temporal mixture-of-experts framework for open-set tactile object recognition

    Long Xiao, Kailin Lyu, Jianwei He, Jianing Zeng, Lin Shu, and Jie Hao. Tacexpert: A pseudo-temporal mixture-of-experts framework for open-set tactile object recognition. InICASSP 2026-2026 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pages 3761–3765. IEEE, 2026

    2026

  22. [22]

    Touch and go: Learning from human-collected vision and touch.arXiv preprint arXiv:2211.12498, 2022

    Fengyu Yang, Chenyang Ma, Jiacheng Zhang, Jing Zhu, Wenzhen Yuan, and Andrew Owens. Touch and go: Learning from human-collected vision and touch.arXiv preprint arXiv:2211.12498, 2022

    Pith/arXiv arXiv 2022

  23. [23]

    Binding touch to everything: Learning unified multimodal tactile representations

    Fengyu Yang, Chao Feng, Ziyang Chen, Hyoungseob Park, Daniel Wang, Yiming Dou, Ziyao Zeng, Xien Chen, Rit Gangopadhyay, Andrew Owens, et al. Binding touch to everything: Learning unified multimodal tactile representations. InProceedings of the IEEE/CVF Confer- ence on Computer Vision and Pattern Recognition, pages 26340–26353, 2024

    2024

  24. [24]

    Vector-quantized image modeling with improved vqgan.arXiv preprint arXiv:2110.04627, 2021

    Jiahui Yu, Xin Li, Jing Yu Koh, Han Zhang, Ruom- ing Pang, James Qin, Alexander Ku, Yuanzhong Xu, Jason Baldridge, and Yonghui Wu. Vector-quantized image modeling with improved vqgan.arXiv preprint arXiv:2110.04627, 2021

    Pith/arXiv arXiv 2021

  25. [25]

    Active clothing material perception using tactile sensing and deep learning

    Wenzhen Yuan, Yuchen Mo, Shaoxiong Wang, and Ed- ward H Adelson. Active clothing material perception using tactile sensing and deep learning. In2018 IEEE International Conference on Robotics and Automation (ICRA), pages 4842–4849. IEEE, 2018

    2018

  26. [26]

    Hardware technology of vision-based tactile sensor: A review.IEEE Sensors Journal, 22(22): 21410–21427, 2022

    Shixin Zhang, Zixi Chen, Yuan Gao, Weiwei Wan, Jian- hua Shan, Hongxiang Xue, Fuchun Sun, Yiyong Yang, and Bin Fang. Hardware technology of vision-based tactile sensor: A review.IEEE Sensors Journal, 22(22): 21410–21427, 2022

    2022

  27. [27]

    Transferable tactile transformers for represen- tation learning across diverse sensors and tasks.arXiv preprint arXiv:2406.13640, 2024

    Jialiang Zhao, Yuxiang Ma, Lirui Wang, and Edward H Adelson. Transferable tactile transformers for represen- tation learning across diverse sensors and tasks.arXiv preprint arXiv:2406.13640, 2024

    Pith/arXiv arXiv 2024