Exploiting Ensemble Learning for Cross-View Isolated Sign Language Recognition

Fei Wang; Kun Li; Peng Zou; Yanyan Wei; Yiqi Nie; Yuwei Wang; Zhangling Duan; Zhiliang Wu

REVIEW 3 major objections 7 minor 3 cited by

Exploiting Ensemble Learning for Cross-View Isolated Sign Language Recognition

T0 review · 3 major / 7 minor · reviewed 2026-08-09 · deepseek-v4-flash

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

Pith's one-line read A two-stage ensemble of Video Swin Transformer sizes (Small, Base, Large) raises cross-view sign recognition to 20.29% top-1 on RGB and 24.53% on RGB-D on MM-WLAuslan.

desk verdict A third-place challenge report that shows a real RGB ensemble gain but leaves the RGB-D claim undersupported; worth a quick read for the ablation, but not a strong paper. read the letter →

arxiv 2502.02196 v1 pith:QDZYMTP5 submitted 2025-02-04 cs.CV cs.AI

Fei Wang , Kun Li , Yiqi Nie , Zhangling Duan , Peng Zou , Zhiliang Wu , Yuwei Wang , Yanyan Wei This is my paper · ORCID

classification cs.CVcs.AI

keywords cross-viewsignlanguagerecognitionensemblelearningVideoSwinTransformermulti-modalfusionRGB-DMM-WLAuslanWWW2025challengetop-1accuracy

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper addresses cross-view isolated sign language recognition, where models trained on frontal-view videos must classify signs recorded from left and right angles. It claims that a two-stage ensemble of Video Swin Transformers—Small, Base, and Large, combined first within each modality and then across RGB and depth—improves robustness to viewpoint change. On the MM-WLAuslan benchmark, the ensemble reaches 20.29% top-1 accuracy for RGB and 24.53% for RGB-D, outperforming every single backbone reported (best single: 17.51% on RGB). The solution ranked third in both tracks of the WWW 2025 CV-ISLR challenge. The load-bearing idea is that models of different capacities capture complementary granularities of gesture features, so averaging their predictions generalizes better to unseen viewpoints.

What carries the argument

The central object is the two-stage ensemble of Video Swin Transformer (VST) models, where VST is a hierarchical video transformer whose 3D shifted-window self-attention jointly models spatial and temporal features. The first stage performs single-modal classification ensemble: weighted sums of the Small, Base, and Large VST predictions within the RGB branch and within the depth branch, with weights $0.4$, $0.4$, $0.2$ respectively. The second stage performs multi-modal fusion ensemble: a weighted combination of the RGB and depth branch outputs with weights $0.65$ and $0.35$. The aggregation is a fixed weighted average of model outputs, not a learned fusion, which keeps the method simple and directly shows the effect of model-capacity diversity plus cross-modal complementarity.

What would settle it

Re-score or retrain the ensemble over a grid of weights—for example, $\lambda_l$ and $\lambda_b$ from 0 to 1 in steps of 0.1 with $\lambda_s = 1-\lambda_l-\lambda_b$, and $\lambda_r$ from 0 to 1 in steps of 0.1—and check whether the reported 20.29% RGB and 24.53% RGB-D top-1 accuracies lie on a broad plateau or a narrow peak; a narrow peak would indicate the ensemble gain is a weight-selection artifact, not a general property of capacity diversity.

Watch

Extended reading notes

Core claim

The central claim is that ensemble learning, applied at two levels, materially improves cross-view sign language recognition over any single Video Swin Transformer. Within each modality, the outputs of VST-Small, VST-Base, and VST-Large are combined with fixed weights $\lambda_l=0.4$, $\lambda_b=0.4$, $\lambda_s=0.2$; for the RGB-D task, the RGB and depth branch outputs are then combined with weights $\lambda_r=0.65$, $\lambda_d=0.35$. The ensemble yields 20.29% top-1 accuracy on the RGB track and 24.53% on the RGB-D track, compared with 17.51% for the best single RGB model and 17.58% for the best single depth model. The paper attributes this gain to the complementary strengths of models with different dimensions in handling viewpoint variability and gesture complexity.

Load-bearing premise

The claimed improvement from ensembling rests on the specific hand-picked weights (0.4, 0.4, 0.2 for the single-modal stage and 0.65, 0.35 for the fusion stage) that the paper reports without any sensitivity analysis, so the gain over the best single model could be an artifact of weights tuned to the validation view rather than evidence of robust complementarity.

Editorial extensions

If this is right

On the MM-WLAuslan benchmark, capacity diversity alone—without changing training data, loss, or architecture family—yields a 2.78-point top-1 gain on the RGB track over the best single VST model.
The same recipe transfers to depth input, where the single best model reaches 17.58% and the full RGB-D ensemble reaches 24.53%, indicating that depth and RGB carry complementary cues for cross-view recognition.
Because the method is a fixed-weight averaging of off-the-shelf VST backbones, it is a cheap robustness lever: any group with pretrained VST variants can reproduce it without redesigning the model.
The third-place finish in both tracks of the WWW 2025 challenge provides a reproducible baseline; the paper's code release lets future work compare directly against this ensemble.
The large gap to the top-ranked teams (56.87% RGB, 57.97% RGB-D) shows that the ensemble alone is not a complete solution to cross-view sign recognition, but it establishes a lower bound that specialized methods must beat.

Reading between the lines

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

The authors leave implicit that the reported gain is sensitive to the chosen weights; if $\lambda_l=0.4, \lambda_b=0.4, \lambda_s=0.2$ and $\lambda_r=0.65, \lambda_d=0.35$ were tuned on the validation view, the ensemble advantage could shrink on a truly novel viewpoint.
An extension the paper does not explore is replacing the fixed weights with a learned aggregation or a small gating network, which might preserve the robustness gain while adapting to view-specific reliability.
The two-stage ensemble idea could be applied to other modality pairs, such as RGB plus skeletal keypoints or RGB plus optical flow; if the cross-modal complementarity is the driver, similar gains should appear there.
Since the paper reports no single-model RGB-D baseline, a direct test of whether multi-modal fusion itself helps—versus just ensembling more models—would require training individual VST-S/B/L on RGB-D input and comparing their average to the reported 24.53%.

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.

Desk Editor's Note

read the letter

The one thing to know: this is a WWW 2025 challenge report that took third place in both CV-ISLR tracks by averaging predictions from three Video Swin Transformer sizes and then fusing RGB and depth. The only genuinely established result is that the RGB ensemble beats each single RGB model (20.29% vs 17.51% best single). Everything else is presented too quickly or overclaimed.

What's good: the paper is short, honest about the gap to the winners (20% vs 57% and 40%), and the RGB ablation in Table 2 directly supports the narrow ensemble claim. The code link is provided. That is a useful data point for anyone working on cross-view sign language recognition.

The soft spots are real. The most important one is that Table 2 has no single-model RGB-D baseline. The 24.53% RGB-D number could be almost entirely due to adding the depth stream, not to ensembling the three model sizes. Eq. 7 implies individual RGB-D models exist, so leaving them out is a genuine omission, not a cosmetic one. The hand-set weights (0.4/0.4/0.2 and 0.65/0.35) also lack any sensitivity analysis, though that is secondary to the missing baseline. There are also no error bars, and the equations in Section 3.3 don't actually show the weights that are described in the implementation details. The phrase "competitive performance" in the abstract is doing too much work when the accuracy is a third of the winner's.

I don't think the central RGB claim is in doubt. But the RGB-D story is undersupported as written, and the conclusion overgeneralizes. The paper is coherent, not incoherent; it is a thin but honest report with one clear gap.

Who this is for: someone organizing or entering a sign-language challenge will find the ablation useful as a simple baseline. A general CV audience won't learn anything new. As a peer-reviewed paper, it needs the missing baseline and a sensitivity analysis before the RGB-D claim is supportable.

My recommendation: send it to peer review only if the venue tolerates workshop-style challenge reports and the authors add the missing RGB-D baselines. As is, a serious editor could either desk reject for low novelty or send it out for a quick revision. I'd lean toward sending it out because the missing experiment is cheap and the question is empirically testable.

Referee Report

3 major / 7 minor

Summary. This paper describes the authors' third-place solution to the WWW 2025 Cross-View Isolated Sign Language Recognition (CV-ISLR) challenge on the MM-WLAuslan dataset. The method trains Video Swin Transformer models in three sizes (Small, Base, Large) on RGB and depth videos, combines the outputs within each modality by a weighted sum of predictions, and then fuses the RGB and depth branches by another weighted sum. The paper reports top-1 accuracies of 20.29% for the RGB track and 24.53% for the RGB-D track, compared with a best single RGB model accuracy of 17.51%. The central claim is that the ensemble strategy captures complementary strengths of models with different capacities and improves cross-view robustness.

Significance. If the reported gains are robust, the paper provides a useful empirical data point on ensembling Video Swin Transformers for cross-view sign language recognition, and the public code repository is a practical strength for reproducibility. The RGB-track result is internally consistent: the ensemble of VST-Small/Base/Large improves over every single RGB model. However, the novelty is limited to an engineering combination of existing components, and the experimental evidence is incomplete for the RGB-D track and for the stability of the hand-set ensemble weights, so the significance as currently presented is modest.

major comments (3)

[Section 4.2, Table 2] The central claim in Section 4.2 that 'the ensemble learning strategy effectively captures the strengths of models with different dimensions' is directly supported only for the RGB track. Table 2 reports VST-Small/Base/Large accuracies for RGB (14.84%/17.51%/17.04%) and for depth (14.01%/16.46%/17.58%), but for the RGB-D track it reports only the ensemble accuracy (24.53%) and leaves the corresponding single-model entries empty. Since Eq. (7) defines the RGB-D ensemble as a combination over model sizes of fused RGB-D models, individual fused models must exist; without reporting at least one single RGB-D baseline, the 4.24-point gap between the RGB ensemble (20.29%) and the RGB-D ensemble (24.53%) could be explained entirely by the additional depth modality rather than by ensembling. Please add per-size fused-model results, or at minimum a VST-Large RGB-D baseline.
[Section 4.1, Implementation Details] The ensemble weights are hand-set (lambda_l:lambda_b:lambda_s = 0.4:0.4:0.2 and lambda_r:lambda_d = 0.65:0.35) with no sensitivity analysis or selection criterion. The reported RGB improvement (20.29% vs. 17.51%) compares a weighted combination against individual models; if these weights were tuned on the validation set, the gain could reflect weight selection rather than a robust property of ensembling. Please report a small sweep over weights (e.g., uniform averaging, majority vote, and neighboring weight vectors) or otherwise justify that the result is insensitive to the choice.
[Section 4.1 / Table 2] No repeated runs, seeds, or error bars are reported in Table 2. The differences among single models are small (VST-Base 17.51% vs. VST-Large 17.04%), and the ensemble gain over VST-Base is 2.78 percentage points; without an estimate of run-to-run variance it is difficult to assess whether this gain is statistically meaningful. If the challenge test set allows only a single evaluation, please state this explicitly and report validation-set results with mean and standard deviation over seeds for the main configurations.

minor comments (7)

[Section 3.3.1, Eqs. (5)-(6)] The summation index i does not appear in the summands of Eqs. (5) and (6); as written, each term is identical. The intended meaning is presumably Phi_i(Z_r) and Phi_i(Z_d) for i in {s,b,l}. Please correct the notation.
[Section 3.3.2, Eq. (7)] Eq. (7) also contains an undefined symbol omega^(i)_rgbd and a summand Phi(Z_r oplus Z_d) that does not depend on i. Please clarify how the fused RGB-D models are indexed and how their parameters enter the sum.
[Section 4.1, Implementation Details] The sentence 'For the VST model, Large, Base and Small sizes of C correspond to 96, 128, and 192' appears to reverse the usual mapping (Large has the largest C). Please verify and state the actual channel widths used for each model size, since the multi-dimensionality claim depends on these settings.
[Section 3.1, Task Definition] The depth input is described as Z_d in R^{T x H x W x 3}; since depth maps are single-channel, please clarify how the three channels are formed (e.g., replication or colorization).
[Table 2] The missing entries in Table 2 (depth ensemble and RGB-D per-backbone rows) should at least be marked with a dash and explained in the caption, so the reader can see which configurations were not evaluated.
[Abstract / Section 4.2] The term 'competitive performance' is not qualified: with a gap of more than 36 points to the first-place RGB result, the claim should be explicitly limited to ranking among submitted challenge systems rather than implying accuracy parity with state-of-the-art methods.
[References] References [22] and [23] are the same paper (Auslan-Daily) listed twice with different years; please deduplicate and use the correct venue/year.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the central claim is an empirical benchmark comparison, not a derivation, and no result reduces to its inputs by construction.

full rationale

The paper's central claim, that ensemble learning improves cross-view sign recognition, is supported by a direct experimental comparison in Table 2: the RGB ensemble reaches 20.29% versus 17.51% for the best single RGB model. This is an empirical measurement on the challenge test set rather than a conclusion derived from the method's equations. Equations (5)-(7) define the ensemble aggregation, but no accuracy value is claimed to follow from those equations without training and testing; no parameter's definition presupposes the target result. The ensemble weights (lambda_l:lambda_b:lambda_s = 0.4:0.4:0.2 and lambda_r:lambda_d = 0.65:0.35) are hand-specified and lack sensitivity analysis, but the paper does not frame them as fitted quantities or as predictions, and the reported 20.29% is not forced by construction. The absence of a single RGB-D baseline in Table 2 is an evidential gap for the multimodal track, but it is a missing comparison, not a circular reduction. The method uses the external Video Swin Transformer backbone [13] and the externally collected MM-WLAuslan dataset [20], neither of which is defined in terms of the paper's ensemble outcome. No load-bearing argument reduces to a self-citation. Accordingly, no circularity is present.

Assumptions & free parameters 2 free parameters · 3 assumptions · 0 invented entities

The paper introduces no new model, entity, or conserved quantity; it relies on standard architectures and hand-set scalar weights. The only free parameters are the ensemble mixing coefficients, which are not justified beyond intuition.

free parameters (2)

Single-modal ensemble weights (lambda_l, lambda_b, lambda_s) = 0.4, 0.4, 0.2
Hand-chosen aggregation weights for VST Large, Base, Small outputs in Eqs. 5-6; no sensitivity analysis is reported.
Multi-modal fusion weights (lambda_r, lambda_d) = 0.65, 0.35
Hand-chosen weights for combining RGB and depth branches in Eq. 7; no sensitivity analysis is reported.

assumptions (3)

domain assumption The MM-WLAuslan challenge test set (left/right views) is a valid measure of cross-view generalization.
The paper's entire evaluation rests on official challenge numbers; no independent validation or protocol details are provided.
domain assumption Video Swin Transformer serves as an adequate baseline for sign language recognition.
The paper adopts VST without comparing to other backbones or justifying its suitability for ISLR beyond citing prior work.
ad hoc to paper Weighted sum of classifier logits is an effective ensemble aggregation method.
No theoretical or empirical comparison to other aggregation methods (e.g., voting, stacking, learned fusion) is given.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Exploiting Ensemble Learning for Cross-View Isolated Sign Language Recognition." pith.science (2026). https://pith.science/paper/QDZYMTP5

@misc{pith2026250202196,
  author       = {Pith},
  title        = {Pith review of: Exploiting Ensemble Learning for Cross-View Isolated Sign Language Recognition},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/QDZYMTP5}},
  note         = {Machine review of arXiv:2502.02196}
}

read the original abstract

In this paper, we present our solution to the Cross-View Isolated Sign Language Recognition (CV-ISLR) challenge held at WWW 2025. CV-ISLR addresses a critical issue in traditional Isolated Sign Language Recognition (ISLR), where existing datasets predominantly capture sign language videos from a frontal perspective, while real-world camera angles often vary. To accurately recognize sign language from different viewpoints, models must be capable of understanding gestures from multiple angles, making cross-view recognition challenging. To address this, we explore the advantages of ensemble learning, which enhances model robustness and generalization across diverse views. Our approach, built on a multi-dimensional Video Swin Transformer model, leverages this ensemble strategy to achieve competitive performance. Finally, our solution ranked 3rd in both the RGB-based ISLR and RGB-D-based ISLR tracks, demonstrating the effectiveness in handling the challenges of cross-view recognition. The code is available at: https://github.com/Jiafei127/CV_ISLR_WWW2025.

Figures

Figures reproduced from arXiv: 2502.02196 by the authors.

**Figure 2.** Illustration of the ensemble learning process. Mul [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗

**Figure 3.** Overview of the proposed architecture for CV-ISLR. The architecture processes RGB and depth videos through Video [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 3 Pith papers

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

Task-Generalized Adaptive Cross-Domain Learning for Multimodal Image Fusion
cs.CV 2025-08 conditional novelty 4.0 of 10

AdaSFFuse combines a learnable wavelet transform and a spatial-frequency Mamba block to report state-of-the-art fusion scores on infrared-visible, multi-exposure, multi-focus, and medical image pairs.
MM-Gesture: Towards Precise Micro-Gesture Recognition through Multimodal Fusion
cs.CV 2025-07 conditional novelty 4.0 of 10

Combining joint, limb, RGB, Taylor-video, optical-flow, and depth streams with two video backbones and a validation-tuned weighted ensemble reaches 73.213% top-1 accuracy on iMiGUE, the best MiGA challenge result to date.
Online Micro-gesture Recognition Using Data Augmentation and Spatial-Temporal Attention
cs.CV 2025-07 reject novelty 3.0 of 10

The paper claims a first-place micro-gesture detection result from data augmentation and spatial-temporal attention, but its own table shows the winning F1 comes from the unmodified AdaTAD baseline, while the proposed...

Reference graph

Works this paper leans on

47 extracted references · 33 canonical work pages · cited by 3 Pith papers

[1]

Yue Cao, Thomas Andrew Geddes, Jean Yee Hwa Yang, and Pengyi Yang. 2020. Ensemble deep learning in bioinformatics. Nature Machine Intelligence 2, 9 (2020), 500–508

work page 2020
[2]

Guoliang Chen, Fei Wang, Kun Li, Zhiliang Wu, Hehe Fan, Yi Yang, Meng Wang, and Dan Guo. 2024. Prototype Learning for Micro-gesture Classification. arXiv preprint arXiv:2408.03097 (2024)

arXiv 2024
[3]

Dan Hendrycks and Kevin Gimpel. 2016. Gaussian error linear units (gelus). arXiv preprint arXiv:1606.08415 (2016)

arXiv 2016
[4]

Dongxu Li, Cristian Rodriguez, Xin Yu, and Hongdong Li. 2020. Word-level Deep Sign Language Recognition from Video: A New Large-scale Dataset and Methods Comparison. In The IEEE Winter Conference on Applications of Computer Vision . 1459–1469

work page 2020
[5]

Kun Li, Dan Guo, Guoliang Chen, Chunxiao Fan, Jingyuan Xu, Zhiliang Wu, Hehe Fan, and Meng Wang. 2024. Prototypical Calibrating Ambiguous Samples for Micro-Action Recognition. arXiv preprint arXiv:2412.14719 (2024)

arXiv 2024
[6]

Kun Li, Dan Guo, Guoliang Chen, Feiyang Liu, and Meng Wang. 2023. Data Augmentation for Human Behavior Analysis in Multi-Person Conversations. In Proceedings of the 31st ACM International Conference on Multimedia . 9516–9520

work page 2023
[7]

Kun Li, Dan Guo, Guoliang Chen, Xinge Peng, and Meng Wang. 2023. Joint skeletal and semantic embedding loss for micro-gesture classification. arXiv preprint arXiv:2307.10624 (2023)

arXiv 2023
[8]

Kun Li, Xinge Peng, Dan Guo, Xun Yang, and Meng Wang. 2024. Repetitive Action Counting with Hybrid Temporal Relation Modeling. arXiv preprint arXiv:2412.07233 (2024)

work page Pith review arXiv 2024

Show all 47 references

[9]

Zhangbin Li, Jinxing Zhou, Jing Zhang, Shengeng Tang, Kun Li, and Dan Guo
[10]

Pengyu Liu, Fei Wang, Kun Li, Guoliang Chen, Yanyan Wei, Shengeng Tang, Zhiliang Wu, and Dan Guo. 2024. Micro-gesture Online Recognition using Learnable Query Points. arXiv preprint arXiv:2407.04490 (2024)

2024 arXiv
[11]

Xiaolong Liu, Qimeng Wang, Yao Hu, Xu Tang, Shiwei Zhang, Song Bai, and Xiang Bai. 2022. End-to-end temporal action detection with transformer. IEEE Transactions on Image Processing 31 (2022), 5427–5441

2022
[12]

Ze Liu, Yutong Lin, Yue Cao, Han Hu, Yixuan Wei, Zheng Zhang, Stephen Lin, and Baining Guo. 2021. Swin transformer: Hierarchical vision transformer us- ing shifted windows. In Proceedings of the IEEE/CVF international conference on computer vision. 10012–10022

2021
[13]

Ze Liu, Jia Ning, Yue Cao, Yixuan Wei, Zheng Zhang, Stephen Lin, and Han Hu. 2022. Video swin transformer. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition . 3202–3211

2022
[14]

Anqi Mao, Mehryar Mohri, and Yutao Zhong. 2023. Cross-entropy loss functions: Theoretical analysis and applications. In International conference on Machine learning. PMLR, 23803–23828

2023
[15]

Wei Qian, Kun Li, Dan Guo, Bin Hu, and Meng Wang. 2024. Cluster-phys: Facial clues clustering towards efficient remote physiological measurement. In Proceedings of the 32nd ACM International Conference on Multimedia . 330–339

2024
[16]

Sneha Sharma, Rinki Gupta, and Arun Kumar. 2022. Trbaggboost: An ensemble- based transfer learning method applied to Indian Sign Language recognition. Journal of Ambient Intelligence and Humanized Computing (2022), 1–11

2022
[17]

Hao Shen, Henghui Ding, Yulun Zhang, Xiaofeng Cong, Zhong-Qiu Zhao, and Xudong Jiang. 2024. Spatial-Frequency Adaptive Remote Sensing Image Dehazing With Mixture of Experts. IEEE Transactions on Geoscience and Remote Sensing (2024)

2024
[18]

Hao Shen, Zhong-Qiu Zhao, and Wandi Zhang. 2023. Adaptive dynamic filtering network for image denoising. In Proceedings of the AAAI Conference on Artificial Intelligence, Vol. 37. 2227–2235

2023
[19]

Hao Shen, Zhong-Qiu Zhao, Yulun Zhang, and Zhao Zhang. 2023. Mutual information-driven triple interaction network for efficient image dehazing. In Proceedings of the 31st ACM International Conference on Multimedia . 7–16

2023
[20]

Xin Shen, Heming Du, Hongwei Sheng, Shuyun Wang, Hui Chen, Huiqiang Chen, Zhuojie Wu, Xiaobiao Du, Jiaying Ying, Ruihan Lu, et al. 2024. MM-WLAuslan: Multi-View Multi-Modal Word-Level Australian Sign Language Recognition Dataset. arXiv preprint arXiv:2410.19488 (2024)

2024 arXiv
[21]

Xin Shen, Lei Shen, Shaozu Yuan, Heming Du, Haiyang Sun, and Xin Yu. 2024. Diverse Sign Language Translation. arXiv preprint arXiv:2410.19586 (2024)

2024 arXiv
[22]

Xin Shen, Shaozu Yuan, Hongwei Sheng, Heming Du, and Xin Yu. 2023. Auslan- Daily: Australian Sign Language Translation for Daily Communication and News. In Advances in Neural Information Processing Systems 36: Annual Conference on Neural Information Processing Systems 2023

2023
[23]

Xin Shen, Shaozu Yuan, Hongwei Sheng, Heming Du, and Xin Yu. 2024. Auslan- daily: Australian sign language translation for daily communication and news. Advances in Neural Information Processing Systems 36 (2024)

2024
[24]

Hongwei Sheng, Xin Shen, Heming Du, Hu Zhang, Zi Huang, and Xin Yu. 2024. AI empowered Auslan learning for parents of deaf children and children of deaf adults. AI and Ethics (2024), 1–11

2024
[25]

Liang Shi, Yixin Chen, Meimei Liu, and Feng Guo. 2024. DuST: Dual Swin Transformer for Multi-modal Video and Time-Series Modeling. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition . 4537–4546

2024
[26]

Shengeng Tang, Dan Guo, Richang Hong, and Meng Wang. 2021. Graph-based multimodal sequential embedding for sign language translation. IEEE Transac- tions on Multimedia 24 (2021), 4433–4445

2021
[27]

Shengeng Tang, Richang Hong, Dan Guo, and Meng Wang. 2022. Gloss semantic- enhanced network with online back-translation for sign language production. In Proceedings of the 30th ACM International Conference on Multimedia . 5630–5638

2022
[28]

Ilya O Tolstikhin, Neil Houlsby, Alexander Kolesnikov, Lucas Beyer, Xiaohua Zhai, Thomas Unterthiner, Jessica Yung, Andreas Steiner, Daniel Keysers, Jakob Uszkoreit, et al. 2021. Mlp-mixer: An all-mlp architecture for vision. Advances in neural information processing systems 3...

2021
[29]

A Vaswani. 2017. Attention is all you need. Advances in Neural Information Processing Systems (2017)

2017
[30]

Fei Wang, Dan Guo, Kun Li, and Meng Wang. 2024. Eulermormer: Robust eulerian motion magnification via dynamic filtering within transformer. In Proceedings of the AAAI Conference on Artificial Intelligence , Vol. 38. 5345–5353

2024
[31]

Fei Wang, Dan Guo, Kun Li, Zhun Zhong, and Meng Wang. 2024. Frequency decoupling for motion magnification via multi-level isomorphic architecture. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 18984–18994

2024
[32]

Xu Wang, Shengeng Tang, Peipei Song, Shuo Wang, Dan Guo, and Richang Hong
[33]

Yu Wang, Fei Wang, Kun Li, Xuping Feng, Wenhui Hou, Lu Liu, Liqing Chen, Yong He, and Yuwei Wang. 2024. Low-light wheat image enhancement using an explicit inter-channel sparse transformer.Computers and Electronics in Agriculture 224 (2024), 109169

2024
[34]

arXiv preprint arXiv:2412.16944 (2024)

Linguistics-Vision Monotonic Consistent Network for Sign Language Production. arXiv preprint arXiv:2412.16944 (2024)

2024 arXiv
[35]

Yanyan Wei, Yilin Zhang, Kun Li, Fei Wang, Shengeng Tang, and Zhao Zhang
[36]

Yiwei Wei, Shaozu Yuan, Meng Chen, Xin Shen, Longbiao Wang, Lei Shen, and Zhiling Yan. 2023. MPP-net: multi-perspective perception network for dense video captioning. Neurocomputing 552 (2023), 126523

2023
[37]

Jiazhen Zhang, Kun Li, Yanyan Wei, Fei Wang, Wei Qian, Jinxing Zhou, and Dan Guo. 2024. Repetitive Action Counting with Feature Interaction Enhancement and Adaptive Gate Fusion. In Proceedings of the 6th ACM International Conference on Multimedia in Asia . 1–7

2024
[38]

Zhenqiang Zhang, Kun Li, Shengeng Tang, Yanyan Wei, Fei Wang, Jinxing Zhou, and Dan Guo. 2025. Temporal Boundary Awareness Network for Repetitive Action Counting. ACM Transactions on Multimedia Computing, Communications and Applications (2025), 1551–6857

2025
[39]

Xiangchen Yin, Donglin Di, Lei Fan, Hao Li, Chen Wei, Xiaofei Gou, Yang Song, Xiao Sun, and Xun Yang. 2024. Grpose: Learning graph relations for human image generation with pose priors. arXiv preprint arXiv:2408.16540 (2024)

2024 arXiv
[40]

Jinxing Zhou, Dan Guo, and Meng Wang. 2022. Contrastive positive sample propagation along the audio-visual event line. IEEE Transactions on Pattern Analysis and Machine Intelligence 45, 6 (2022), 7239–7257

2022
[41]

Jinxing Zhou, Jianyuan Wang, Jiayi Zhang, Weixuan Sun, Jing Zhang, Stan Birchfield, Dan Guo, Lingpeng Kong, Meng Wang, and Yiran Zhong. 2022. Audio– visual segmentation. In European Conference on Computer Vision . Springer, 386– 403

2022
[42]

Jiaqi Zhao, Fei Wang, Kun Li, Yanyan Wei, Shengeng Tang, Shu Zhao, and Xiao Sun. 2024. Temporal-Frequency State Space Duality: An Efficient Paradigm for Speech Emotion Recognition. arXiv preprint arXiv:2412.16904 (2024)

2024 arXiv
[43]

Xiaojun Zhou, Jingyi He, and Chunhua Yang. 2022. An ensemble learning method based on deep neural network and group decision making. Knowledge-Based Conference acronym ’XX, June 03–05, 2018, Woodstock, NY Fei Wang et al. Systems 239 (2022), 107801

2022
[44]

Ziheng Zhou, Jinxing Zhou, Wei Qian, Shengeng Tang, Xiaojun Chang, and Dan Guo. 2024. Dense Audio-Visual Event Localization under Cross-Modal Consistency and Multi-Temporal Granularity Collaboration. arXiv preprint arXiv:2412.12628 (2024)

2024 arXiv
[45]

Pan Zhou, Xingyu Xie, Zhouchen Lin, and Shuicheng Yan. 2024. Towards un- derstanding convergence and generalization of AdamW. IEEE Transactions on Pattern Analysis and Machine Intelligence (2024)

2024
[2024]

arXiv preprint arXiv:2412.10749 (2024)

Patch-level Sounding Object Tracking for Audio-Visual Question Answer- ing. arXiv preprint arXiv:2412.10749 (2024)

2024 arXiv
[2025]

Computer Vision and Image Understanding 250 (2025), 104222

Leveraging vision-language prompts for real-world image restoration and enhancement. Computer Vision and Image Understanding 250 (2025), 104222

2025

Pith tools

Reviewed August 9, 2026 · model on record in the stance chip above.

[1] [1]

Yue Cao, Thomas Andrew Geddes, Jean Yee Hwa Yang, and Pengyi Yang. 2020. Ensemble deep learning in bioinformatics. Nature Machine Intelligence 2, 9 (2020), 500–508

work page 2020

[2] [2]

Guoliang Chen, Fei Wang, Kun Li, Zhiliang Wu, Hehe Fan, Yi Yang, Meng Wang, and Dan Guo. 2024. Prototype Learning for Micro-gesture Classification. arXiv preprint arXiv:2408.03097 (2024)

arXiv 2024

[3] [3]

Dan Hendrycks and Kevin Gimpel. 2016. Gaussian error linear units (gelus). arXiv preprint arXiv:1606.08415 (2016)

arXiv 2016

[4] [4]

Dongxu Li, Cristian Rodriguez, Xin Yu, and Hongdong Li. 2020. Word-level Deep Sign Language Recognition from Video: A New Large-scale Dataset and Methods Comparison. In The IEEE Winter Conference on Applications of Computer Vision . 1459–1469

work page 2020

[5] [5]

Kun Li, Dan Guo, Guoliang Chen, Chunxiao Fan, Jingyuan Xu, Zhiliang Wu, Hehe Fan, and Meng Wang. 2024. Prototypical Calibrating Ambiguous Samples for Micro-Action Recognition. arXiv preprint arXiv:2412.14719 (2024)

arXiv 2024

[6] [6]

Kun Li, Dan Guo, Guoliang Chen, Feiyang Liu, and Meng Wang. 2023. Data Augmentation for Human Behavior Analysis in Multi-Person Conversations. In Proceedings of the 31st ACM International Conference on Multimedia . 9516–9520

work page 2023

[7] [7]

Kun Li, Dan Guo, Guoliang Chen, Xinge Peng, and Meng Wang. 2023. Joint skeletal and semantic embedding loss for micro-gesture classification. arXiv preprint arXiv:2307.10624 (2023)

arXiv 2023

[8] [8]

Kun Li, Xinge Peng, Dan Guo, Xun Yang, and Meng Wang. 2024. Repetitive Action Counting with Hybrid Temporal Relation Modeling. arXiv preprint arXiv:2412.07233 (2024)

work page Pith review arXiv 2024