Pith. sign in

REVIEW 4 major objections 5 minor 35 references

ConViTac: Aligning Visual-Tactile Fusion with Contrastive Representations

T0 review · 4 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read ConViTac claims that conditioning visual-tactile fusion on frozen contrastive embeddings through cross-modal attention aligns the two modalities and improves accuracy by up to 12.0% on material classification and grasping prediction.

desk verdict A useful visual-tactile fusion recipe whose headline gains are confounded by the added attention module; worth a serious referee, but the attribution needs an unconfounded control. read the letter →

arxiv 2506.20757 v1 pith:DNWT4CZS submitted 2025-06-25 cs.CV cs.RO

Zhiyuan Wu , Yongqiang Zhao , Shan Luo This is my paper
classification cs.CVcs.RO
keywords visual-tactilefusioncontrastivelearningcross-modalattentionmaterialclassificationgraspingpredictionself-supervisedpretrainingroboticperceptionmultimodalrepresentation
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

ConViTac is a visual-tactile representation learning network whose central claim is that feature fusion works better when it is conditioned on contrastive embeddings that already live in a shared vision-touch space. The paper argues that direct fusion, adding or concatenating visual and tactile features, leaves the modalities poorly associated, whereas using pretrained contrastive embeddings to guide a cross-modal attention module aligns the features before downstream classification. If this is right, a frozen self-supervised contrastive encoder becomes a reusable asset for supervised robotic perception, injecting cross-modal correspondence into tasks like material classification and grasp-success prediction without being retrained. The paper reports accuracy gains up to 12.0% over strong baselines on Touch and Go and ObjectFolder Real, plus 84.3% grasp prediction accuracy on The Feeling of Success dataset.

What carries the argument

The load-bearing component is the Contrastive Embedding Conditioning (CEC) mechanism. It has two parts: a frozen contrastive encoder $E_c$, chosen as DINO, pretrained with a SimCLR-style contrastive loss on visual-tactile data, which projects each visual frame and tactile frame into a shared latent space and concatenates them into $e_p$; and a cross-modal attention module in which $e_p$ serves as the query while the fused feature map $f_f$ supplies keys and values, producing the aligned feature $f_a$. The mechanism's job is to use the contrastive embeddings as a condition that tells the attention where visual and tactile features correspond, aligning the feature distributions before the supervised head reads them.

What would settle it

Retrain ConViTac in two controlled variants: one with the frozen contrastive embeddings replaced by random projections of the same dimension, and one with the contrastive encoder pretrained only on the training split, then compare accuracy on Touch and Go, ObjectFolder Real, and The Feeling of Success. If either variant keeps the reported gains, or if excluding test frames removes them, the central claim is false.

Watch

Extended reading notes

Core claim

The paper's discovery, stated on its own terms, is that the alignment problem in visual-tactile fusion can be treated as a conditioning problem. After a contrastive encoder is pretrained in a self-supervised way with a SimCLR-style loss, it is frozen and used to project every visual and tactile image into a unified latent space. The concatenation of these projected embeddings is then used as the query in a multi-head cross-modal attention module whose keys and values come from the fused visual-tactile feature map. This operation reshapes the fused representation so that visually and tactically corresponding regions reinforce each other. In experiments, this Contrastive Embedding Conditioning mechanism improves all tested fusion strategies, and the full ConViTac network exceeds previous contrastive and supervised visual-tactile methods on material classification and grasping-success prediction.

Load-bearing premise

The result rests on the assumption that the accuracy gains come from the contrastive alignment signal itself, and not from the extra cross-modal attention parameters or from contrastive pretraining that may have seen test data.

Editorial extensions

If this is right

  • Frozen contrastive embeddings can be reused across downstream supervised tasks, so only the fusion and task head need to be trained for each new task.
  • CEC is modular over fusion operators: addition, concatenation, and softmax weighted sum all improve when conditioned, by 3.3 to 9.2 accuracy points.
  • Contrastive alignment and cross-entropy supervision are complementary rather than competing, since the combined network beats both pure contrastive linear-probe methods and pure supervised direct-fusion baselines.
  • The added cost, 91.79 MiB of parameters and a 16.6% speed reduction, still leaves the network at 31.85 FPS, so the mechanism remains practical for real-time robotic use.
  • Qualitative evidence indicates the mechanism shifts attention toward contact and grasp regions, and PCA visualization shows the visual and tactile feature distributions becoming more aligned, which supports the alignment story behind the accuracy gains.

Reading between the lines

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

  • The same conditioning recipe likely transfers to other paired image modalities beyond vision and touch, such as RGB-depth or vision-audio, because the mechanism only requires a shared latent space plus cross-modal attention.
  • A parameter-matched control is the decisive missing test: since the ablations add CEC, contrastive encoder plus attention, to baselines that have neither, part of the reported gain could be extra capacity rather than the alignment signal.
  • If the contrastive embeddings genuinely focus attention on contact regions, they could serve as weak supervision for contact localization or affordance prediction, tasks the paper does not evaluate.
  • The paper's stated future directions, peg insertion and lock opening, suggest the natural next check is whether the alignment benefit persists in closed-loop manipulation, where action and contact feedback matter as much as static labels.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 5 minor

Summary. The paper proposes ConViTac, a visual-tactile fusion network whose central component is a Contrastive Embedding Conditioning (CEC) mechanism. A contrastive encoder is first pretrained (the text says via SimCLR, and later says DINO is used) to project visual and tactile images into a shared latent space; these projected embeddings are then fed as queries into a cross-modal attention module that conditions the fusion of visual and tactile features. The method is evaluated on material classification (Touch and Go, ObjectFolder Real) and grasp-success prediction (Feeling of Success), with reported improvements over existing supervised and contrastive baselines of up to 12.0 percentage points. The authors also present ablations over fusion operations, conditioning modalities, and encoder architectures, plus Grad-CAM and PCA visualizations.

Significance. If the central claim holds, the paper offers a simple and transferable recipe: use frozen contrastively aligned embeddings to condition cross-modal attention rather than merely concatenating features. The evaluation is on real-world benchmarks and covers a reasonable set of baselines, fusion operators, and qualitative analyses. The paper is also honest about computational overhead, reporting parameter and FPS changes. However, the significance is conditional on three unresolved points: the attribution of the gains to contrastive alignment rather than added attention capacity, the possibility of test-data contamination in the self-supervised pretraining stage, and the technical inconsistencies about what the contrastive encoder actually is (SimCLR vs. DINO). Because these points affect the interpretation of the headline 12.0% improvement, the empirical contribution is not yet fully established.

major comments (4)
  1. [III-C.1 and Fig. 2] The pretraining stage is not specified clearly enough to rule out test-set leakage. The text states that Ec is trained "within all visual-tactile data" using self-supervised contrastive learning, but it does not confirm that the evaluation splits of Touch and Go, ObjectFolder Real, and Feeling of Success are excluded from this stage. Since the downstream classifiers and the CEC conditioning both use representations produced by Ec, including test frames in pretraining could inflate the reported gains, including the headline 12.0% improvement. The authors should state explicitly which data splits are used for training Ec, and if "all visual-tactile data" literally includes test frames, the experiments must be rerun on a leakage-free protocol.
  2. [IV-C.2, IV-C.3, Tables IV and V] The ablation studies do not isolate the effect of contrastive alignment from the effect of adding a cross-modal attention module. In Tables IV and V, every "-Con" variant adds both the pretrained contrastive encoder Ec and the multi-head cross-modal attention of Eqs. (5)-(7) on top of a baseline that has neither; there is no control that adds the attention module with a non-contrastive or random conditioning input. The small increments of DINO over ViT in Table III (2.1, 0.2, and 3.4 points on the three datasets) suggest that much of the improvement over the unconditional baseline may come from attention capacity rather than from the contrastive alignment signal. The authors should add controls such as a randomly initialized frozen encoder, a non-contrastive learned encoder, or a constant learnable query fed into the same attention module.
  3. [III-C.1, Eq. (3), and reference [30]] There is a technical inconsistency in the description of the contrastive encoder. The loss in Eq. (3) is the SimCLR/NT-Xent contrastive loss, and the text says the encoder is trained "through SimCLR [16]", but immediately afterward it states "we utilize DINO [30] as Ec". SimCLR and DINO are different self-supervised methods, and cited reference [30] is actually "DINO: DETR with Improved DeNoising anchor boxes", an object-detection method, not the self-supervised DINO of Caron et al. The authors must clarify whether Ec is trained with Eq. (3), whether it is an off-the-shelf DINO model, or whether it is a DINO-initialized model fine-tuned with Eq. (3), and they must supply the correct citation. This matters because the paper's central claim is that a contrastively trained encoder produces the alignment benefit.
  4. [Tables I-V] All accuracy results are reported as single numbers with no variance or significance measures. Given that some key comparisons are small (e.g., DINO vs. ViT on the Feeling of Success dataset is a 0.4-point difference, and several Table IV conditioning-modality differences are between 1.3 and 4.9 points), the central empirical claims need standard deviations over multiple training seeds or paired significance tests. Otherwise the improvements attributed to CEC may not be statistically distinguishable from training noise.
minor comments (5)
  1. [IV-C.1] The phrase "Contrastive Encoder Component (CEC)" should be "Contrastive Embedding Conditioning" to match the acronym introduced in the title and Section III-C.
  2. [Fig. 1 and V] The caption of Fig. 1 uses "substream tasks"; this should be "downstream tasks".
  3. [Eq. (1)] The description "concatenate F_v along dimension 0" is ambiguous because F_v is already written as R^{N x P x D} and the resulting feature map is stated to be R^{2N x P x D}; please clarify which dimension is being concatenated and how the linear projections L_v and L_t are applied.
  4. [IV-A] The learning rate of 0.1 with the Adam optimizer is unusually high; if this is intentional, a sentence explaining that it worked well in practice would help reproducibility.
  5. [IV-B.1] The sentence "ConViTac outperforms baseline methods by 33.7% to 11.4%" is unclear about whether these are absolute percentage-point differences or relative improvements; please state the comparison explicitly for each baseline.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: ConViTac's CEC mechanism is an empirical architecture comparison, not a derivation that reduces to its own inputs.

full rationale

The paper's central claim is an empirical one: adding a frozen contrastive encoder and cross-modal attention improves visual-tactile fusion accuracy on downstream tasks. The contrastive encoder is pretrained with a standard self-supervised objective on the same visual-tactile datasets, but the supervised downstream objective (cross-entropy on task labels) is distinct from the contrastive loss, and the reported gains are not mathematically forced by the contrastive embeddings themselves. No equation in Section III sets the final prediction equal to the contrastive embedding by construction, and no fitted parameter is renamed as a prediction; Tables IV and V compare architectural variants rather than deriving results from fitted values. The main caveat is the phrase in Sec. III-C.1 that Ec is trained 'within all visual-tactile data,' which could imply that test frames are included in pretraining; this is a data-leakage or evaluation-validity concern, not circular reasoning, because the downstream labels and classifier are still learned independently. Self-citations in the paper are background or architectural inspiration and are not load-bearing justifications, and no uniqueness theorem is invoked to force the method. Therefore there is no circularity of the kind defined in this review.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The method rests on domain assumptions about synchronized data, the validity of SimCLR on this modality pair, whether pretraining excludes test data, and whether contrastive conditioning is separable from extra attention capacity.

free parameters (3)
  • temperature parameter tau in contrastive loss (Eq. 4) = not specified
    Scales logits in the SimCLR-style loss; the value is chosen by hand and not reported, and it can affect the quality of the contrastive embeddings.
  • number of attention heads h in Eq. 5 = 8
    Empirically set to 8; it affects the capacity and behavior of the cross-modal attention module.
  • patch size P = 16
    Fixed at 16 for the ViT encoders; it controls input granularity and influences the granularity of the fused features.
assumptions (4)
  • domain assumption The visual and tactile sequences are synchronized with one-to-one correspondence, so each frame pair can be treated as a matched multimodal sample.
    Stated in Sec. III-A for the task setup; if the correspondence is imperfect, contrastive alignment would be trained on false positives.
  • domain assumption A SimCLR-style contrastive loss over concatenated visual-tactile projections learns a joint space where matching modality pairs are closer.
    Invoked in Sec. III-C.1 as the basis for pretraining Ec; the paper provides no formal or empirical check that the learned space is actually aligned.
  • domain assumption Self-supervised pretraining on 'all visual-tactile data' does not leak task-relevant information from test splits.
    Sec. III-C.1 says training Ec 'within all visual-tactile data'; the paper does not state whether this is restricted to training splits.
  • ad hoc to paper The improvement from CEC is due to contrastive conditioning rather than the added cross-modal attention parameters.
    Ablations in Sec. IV-C compare no-CEC baselines that lack the attention module, so the two changes are confounded; this assumption is central to the paper's attribution.

how reviews work

0 comments
Cite this review

Pith. "Pith review of ConViTac: Aligning Visual-Tactile Fusion with Contrastive Representations." pith.science (2026). https://pith.science/paper/DNWT4CZS

@misc{pith2026250620757,
  author       = {Pith},
  title        = {Pith review of: ConViTac: Aligning Visual-Tactile Fusion with Contrastive Representations},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/DNWT4CZS}},
  note         = {Machine review of arXiv:2506.20757}
}
read the original abstract

Vision and touch are two fundamental sensory modalities for robots, offering complementary information that enhances perception and manipulation tasks. Previous research has attempted to jointly learn visual-tactile representations to extract more meaningful information. However, these approaches often rely on direct combination, such as feature addition and concatenation, for modality fusion, which tend to result in poor feature integration. In this paper, we propose ConViTac, a visual-tactile representation learning network designed to enhance the alignment of features during fusion using contrastive representations. Our key contribution is a Contrastive Embedding Conditioning (CEC) mechanism that leverages a contrastive encoder pretrained through self-supervised contrastive learning to project visual and tactile inputs into unified latent embeddings. These embeddings are used to couple visual-tactile feature fusion through cross-modal attention, aiming at aligning the unified representations and enhancing performance on downstream tasks. We conduct extensive experiments to demonstrate the superiority of ConViTac in real world over current state-of-the-art methods and the effectiveness of our proposed CEC mechanism, which improves accuracy by up to 12.0% in material classification and grasping prediction tasks.

Figures

Figures reproduced from arXiv: 2506.20757 by the authors.

Figure 1
Figure 1. An example of shovel in ObjectFolder Real [1] for our ConViTac. Given a visual image and a corresponding tactile image, we obtain aligned [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. An overview of our proposed ConViTac architecture. Our ConViTac processes visual and tactile sequences through a two-stage [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Visualization for the effectiveness of our CEC mechanism using Grad-Cam [24] with highlights in red boxes. With CEC mechanism, our ConViTac [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Visualization for visual and tactile feature distribution with [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

35 extracted references · 18 canonical work pages

  1. [30]

    Dino: Detr with improved denoising anchor boxes for end-to-end object detection,

    H. Zhang, F. Li, S. Liu, L. Zhang, H. Su, J. Zhu, L. M. Ni, and H.-Y . Shum, “Dino: Detr with improved denoising anchor boxes for end-to-end object detection,”arXiv preprint arXiv:2203.03605, 2022

    arXiv 2022

  2. [16]

    A simple frame- work for contrastive learning of visual representations,

    T. Chen, S. Kornblith, M. Norouzi, and G. Hinton, “A simple frame- work for contrastive learning of visual representations,” inInterna- tional conference on machine learning. PmLR, 2020, pp. 1597–1607

    2020

  3. [1]

    The objectfolder benchmark: Multisensory learning with neural and real objects,

    R. Gao, Y . Dou, H. Li, T. Agarwal, J. Bohg, Y . Li, L. Fei-Fei, and J. Wu, “The objectfolder benchmark: Multisensory learning with neural and real objects,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2023, pp. 17 276–17 286

    work page 2023

  4. [2]

    Robotic tactile perception of object properties: A review,

    S. Luo, J. Bimbo, R. Dahiya, and H. Liu, “Robotic tactile perception of object properties: A review,”Mechatronics, vol. 48, pp. 54–67, 2017

    2017

  5. [3]

    Deep domain adaptation regression for force calibration of optical tactile sensors,

    Z. Chen, N. Ou, J. Jiang, and S. Luo, “Deep domain adaptation regression for force calibration of optical tactile sensors,” in2024 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE, 2024, pp. 13 561–13 568

    work page 2024

  6. [4]

    The feeling of success: Does touch sensing help predict grasp outcomes?

    R. Calandra, A. Owens, M. Upadhyaya, W. Yuan, J. Lin, E. H. Adelson, and S. Levine, “The feeling of success: Does touch sensing help predict grasp outcomes?” inConference on Robot Learning (CoRL). PMLR, 2017, pp. 314–323

    work page 2017

  7. [5]

    Spatio-temporal attention model for tactile texture recognition,

    G. Cao, Y . Zhou, D. Bollegala, and S. Luo, “Spatio-temporal attention model for tactile texture recognition,” in2020 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE, 2020, pp. 9896–9902

    work page 2020

  8. [6]

    Multimodal visual-tactile rep- resentation learning through self-supervised contrastive pre-training,

    V . Dave, F. Lygerakis, and E. Rueckert, “Multimodal visual-tactile rep- resentation learning through self-supervised contrastive pre-training,” arXiv preprint arXiv:2401.12024, 2024

    arXiv 2024

Show all 35 references
  1. [7]

    Vitac: Feature sharing between vision and tactile sensing for cloth texture recognition,

    S. Luo, W. Yuan, E. Adelson, A. G. Cohn, and R. Fuentes, “Vitac: Feature sharing between vision and tactile sensing for cloth texture recognition,” in2018 IEEE International Conference on Robotics and Automation (ICRA). IEEE, 2018, pp. 2722–2727

    2018

  2. [8]

    Gelsight: High-resolution robot tactile sensors for estimating geometry and force,

    W. Yuan, S. Dong, and E. H. Adelson, “Gelsight: High-resolution robot tactile sensors for estimating geometry and force,”Sensors, vol. 17, no. 12, p. 2762, 2017

    2017

  3. [9]

    Geltip: A finger-shaped optical tac- tile sensor for robotic manipulation,

    D. F. Gomes, Z. Lin, and S. Luo, “Geltip: A finger-shaped optical tac- tile sensor for robotic manipulation,” in2020 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE, 2020, pp. 9903–9909

    2020

  4. [10]

    Touchroller: A rolling optical tactile sensor for rapid assessment of large surfaces,

    G. Cao, J. Jiang, C. Lu, D. F. Gomes, and S. Luo, “Touchroller: A rolling optical tactile sensor for rapid assessment of large surfaces,” arXiv preprint arXiv:2103.00595, 2021

    2021 arXiv

  5. [11]

    Self-attention based visual-tactile fusion learning for predicting grasp outcomes,

    S. Cui, R. Wang, J. Wei, J. Hu, and S. Wang, “Self-attention based visual-tactile fusion learning for predicting grasp outcomes,”IEEE Robotics and Automation Letters, vol. 5, no. 4, pp. 5827–5834, 2020

    2020

  6. [12]

    Touch and go: Learning from human-collected vision and touch,

    F. Yang, C. Ma, J. Zhang, J. Zhu, W. Yuan, and A. Owens, “Touch and go: Learning from human-collected vision and touch,”Advances in Neural Information Processing Systems (NeurIPS), vol. 35, pp. 8081– 8103, 2022

    2022

  7. [13]

    Self-supervised visuo-tactile pretraining to locate and follow garment features,

    J. Kerr, H. Huang, A. Wilcox, R. Hoque, J. Ichnowski, R. Calandra, and K. Goldberg, “Self-supervised visuo-tactile pretraining to locate and follow garment features,”arXiv preprint arXiv:2209.13042, 2022

    2022 arXiv

  8. [14]

    Multisensory integration: How visual experience shapes spatial perception,

    M. Eimer, “Multisensory integration: How visual experience shapes spatial perception,”Current biology, vol. 14, no. 3, pp. R115–R117, 2004

    2004

  9. [15]

    Learning transferable visual models from natural language supervision,

    A. Radford, J. W. Kim, C. Hallacy, A. Ramesh, G. Goh, S. Agarwal, G. Sastry, A. Askell, P. Mishkin, J. Clarket al., “Learning transferable visual models from natural language supervision,” inInternational conference on machine learning (ICML). PMLR, 2021, pp. 8748– 8763

    2021

  10. [17]

    Transforce: Transferable force prediction for vision-based tactile sensors with sequential image translation,

    Z. Chen, N. Ou, X. Zhang, and S. Luo, “Transforce: Transferable force prediction for vision-based tactile sensors with sequential image translation,”arXiv preprint arXiv:2409.09870, 2024

    2024 arXiv

  11. [18]

    Learning continuous grasp stability for a humanoid robot hand based on tactile sensing,

    J. Schill, J. Laaksonen, M. Przybylski, V . Kyrki, T. Asfour, and R. Dillmann, “Learning continuous grasp stability for a humanoid robot hand based on tactile sensing,” in2012 4th IEEE RAS & EMBS International Conference on Biomedical Robotics and Biomechatron- ics (BioRob). I...

    2012

  12. [19]

    Visuo-tactile transformers for manipulation,

    Y . Chen, M. Van der Merwe, A. Sipos, and N. Fazeli, “Visuo-tactile transformers for manipulation,” in6th Annual Conference on Robot Learning, 2022

    2022

  13. [20]

    Contrastive multimodal fusion with tupleinfonce,

    Y . Liu, Q. Fan, S. Zhang, H. Dong, T. Funkhouser, and L. Yi, “Contrastive multimodal fusion with tupleinfonce,” inProceedings of the IEEE/CVF International Conference on Computer Vision, 2021, pp. 754–763

    2021

  14. [21]

    Representation learning with contrastive predictive coding,

    A. v. d. Oord, Y . Li, and O. Vinyals, “Representation learning with contrastive predictive coding,”arXiv preprint arXiv:1807.03748, 2018

    2018 arXiv

  15. [22]

    Momentum contrast for unsupervised visual representation learning,

    K. He, H. Fan, Y . Wu, S. Xie, and R. Girshick, “Momentum contrast for unsupervised visual representation learning,” inProceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2020, pp. 9729–9738

    2020

  16. [23]

    Binding touch to everything: Learning unified multimodal tactile representations,

    F. Yang, C. Feng, Z. Chen, H. Park, D. Wang, Y . Dou, Z. Zeng, X. Chen, R. Gangopadhyay, A. Owenset al., “Binding touch to everything: Learning unified multimodal tactile representations,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 20...

    2024

  17. [24]

    Grad-cam: Visual explanations from deep networks via gradient-based localization,

    R. R. Selvaraju, M. Cogswell, A. Das, R. Vedantam, D. Parikh, and D. Batra, “Grad-cam: Visual explanations from deep networks via gradient-based localization,” inProceedings of the IEEE international conference on computer vision, 2017, pp. 618–626

    2017

  18. [25]

    An image is worth 16x16 words: Transformers for image recognition at scale,

    A. Dosovitskiy, L. Beyer, A. Kolesnikov, D. Weissenborn, X. Zhai, T. Unterthiner, M. Dehghani, M. Minderer, G. Heigold, S. Gellyet al., “An image is worth 16x16 words: Transformers for image recognition at scale,” inInternational Conference on Learning Representations (ICLR), 2020

    2020

  19. [26]

    S 3m- net: Joint learning of semantic segmentation and stereo matching for autonomous driving,

    Z. Wu, Y . Feng, C.-W. Liu, F. Yu, Q. Chen, and R. Fan, “S 3m- net: Joint learning of semantic segmentation and stereo matching for autonomous driving,”IEEE Transactions on Intelligent Vehicles, 2024

    2024

  20. [27]

    Sg- roadseg: End-to-end collision-free space detection sharing encoder representations jointly learned via unsupervised deep stereo,

    Z. Wu, J. Li, Y . Feng, C. Liu, W. Ye, Q. Chen, and R. Fan, “Sg- roadseg: End-to-end collision-free space detection sharing encoder representations jointly learned via unsupervised deep stereo,” in2024 International Conference on Robotics and Automation (ICRA). IEEE, 2024, p. in press

    2024

  21. [28]

    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,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2022, pp. 10 684–10 695

    2022

  22. [29]

    Cdi3d: Cross-guided dense-view interpolation for 3d reconstruction,

    Z. Wu, X. Song, S. Wang, W. Liu, J. Yang, Z. Cheng, S. Chen, T. Shang, W. Sun, S. Luoet al., “Cdi3d: Cross-guided dense-view interpolation for 3d reconstruction,”arXiv preprint arXiv:2503.08005, 2025

    2025 arXiv

  23. [31]

    Adam: A method for stochastic optimiza- tion,

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

    2014 arXiv

  24. [32]

    Deep residual learning for image recognition,

    K. He, X. Zhang, S. Ren, and J. Sun, “Deep residual learning for image recognition,” inProceedings of the IEEE conference on computer vision and pattern recognition, 2016, pp. 770–778

    2016

  25. [33]

    Two-dimensional pca: a new approach to appearance-based face representation and recognition,

    J. Yang, D. Zhang, A. F. Frangi, and J.-y. Yang, “Two-dimensional pca: a new approach to appearance-based face representation and recognition,”IEEE transactions on pattern analysis and machine intelligence, vol. 26, no. 1, pp. 131–137, 2004

    2004

  26. [34]

    Learning deep multimodal feature representation with asymmetric multi-layer fusion,

    Y . Wang, F. Sun, M. Lu, and A. Yao, “Learning deep multimodal feature representation with asymmetric multi-layer fusion,” inPro- ceedings of the 28th ACM International Conference on Multimedia (ACM MM), 2020, pp. 3902–3910

    2020

  27. [35]

    Multimodal zero- shot learning for tactile texture recognition,

    G. Cao, J. Jiang, D. Bollegala, M. Li, and S. Luo, “Multimodal zero- shot learning for tactile texture recognition,”Robotics and Autonomous Systems, vol. 176, p. 104688, 2024

    2024

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.