A Unified Multimodal De- and Re-Coupling Framework for RGB-D Motion Recognition
Open accessThe DOI matches its Crossref record; 86 of the 86 references it cites were found in a registry or an index.
Abstract
Motion recognition is a promising direction in computer vision, but the training of video classification models is much harder than images due to insufficient data and considerable parameters. To get around this, some works strive to explore multimodal cues from RGB-D data. Although improving motion recognition to some extent, these methods still face sub-optimal situations in the following aspects: (i) Data augmentation, i.e., the scale of the RGB-D datasets is still limited, and few efforts have been made to explore novel data augmentation strategies for videos; (ii) Optimization mechanism, i.e., the tightly space-time-entangled network structure brings more challenges to spatiotemporal information modeling; And (iii) cross-modal knowledge fusion, i.e., the high similarity between multimodal representations leads to insufficient late fusion. To alleviate these drawbacks, we propose to improve RGB-D-based motion recognition both from data and algorithm perspectives in this article. In more detail, firstly, we introduce a novel video data augmentation method dubbed ShuffleMix, which acts as a supplement to MixUp, to provide additional temporal regularization for motion recognition. Secondly, a Unified Multimodal De-coupling and multi-stage Re-coupling framework, termed UMDR, is proposed for video representation learning. Finally, a novel cross-modal Complement Feature Catcher (CFCer) is explored to mine potential commonalities features in multimodal information as the auxiliary fusion stream, to improve the late fusion results. The seamless combination of these novel designs forms a robust spatiotemporal representation and achieves better performance than state-of-the-art methods on four public motion datasets. Specifically, UMDR achieves unprecedented improvements of ↑ 4.5% on the Chalearn IsoGD dataset.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“However, existing works on missing modality robustness vary widely in terms of model architectures [32, 4], fusion methods [35, 63], training procedures [20, 32], and missing feature generation methods [23]. Due to this heterogeneity, a fair comparison based solely on model size/number of parameters is infeasible.”
From Robust Multimodal Learning with Missing Modalities via Parameter-Efficient Adaptation (2025) · cites this paper as [63]
Experimental — accuracy not yet measuredThe cited source says what the claim attributes to it.
In the citing paper
“As shown in Table 5, our adaptation performs better compared to recent modality masking and generation based approach ActionMAE [23] and modality de- and re-coupling based approaches Motion-RGBD [91] and UMDR [63].”
From Robust Multimodal Learning with Missing Modalities via Parameter-Efficient Adaptation (2025) · cites this paper as [63]
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale
Dosovitskiy, A., Beyer, L., Kolesnikov, A., Weissenborn, D., Zhai, X., Unterthiner, T., Dehghani, M., Minderer, M., Heigold, G., Gelly, S., et al. (2020) · arXiv · doi:10.48550/arxiv.2010.11929 · Cited by 47,472 · cited 1×
“To achieve the goal in the former aspect, inspired by self-attention [46], we calculate dynamic attention maps based on the input to re-weight spatial features of each frame to highlight sensitive filters in it.”
MethodDistilling the Knowledge in a Neural Network
Hinton, G., Vinyals, O., & Dean, J. (2015) · arXiv · doi:10.48550/arxiv.1503.02531 · Cited by 14,204
A survey on Image Data Augmentation for Deep Learning
Shorten, C., & Khoshgoftaar, T. M. (2019) · Journal of Big Data · doi:10.1186/s40537-019-0197-0 · Cited by 13,203 · cited 1×
“Data augmentation, a data-driven and informed regularization strategy that artificially increases the number of training samples to improve the generalization ability of models, has been widely explored for better training of deep neural networks [20], [21], [22], [23].”
BackgroundQuo Vadis, Action Recognition? A New Model and the Kinetics Dataset
Carreira, J., & Zisserman, A. (2017) · 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2017.502 · Cited by 9,570
mixup: Beyond Empirical Risk Minimization
Zhang, H., Cisse, M., Dauphin, Y. N., & Lopez-Paz, D. (2017) · arXiv · doi:10.48550/arxiv.1710.09412 · Cited by 5,831 · cited 3×
“The latter refers to utilizing multiple samples to generate a new instance, including video MixUp [14] and CutMix [28].”
BackgroundMethodCutMix: Regularization Strategy to Train Strong Classifiers With Localizable Features
Yun, S., Han, D., Chun, S., Oh, S. J., Yoo, Y., & Choe, J. (2019) · 2019 IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2019.00612 · Cited by 5,400 · cited 1×
“The latter refers to utilizing multiple samples to generate a new instance, including video MixUp [14] and CutMix [28].”
BackgroundSGDR: Stochastic Gradient Descent with Warm Restarts
Loshchilov, I., & Hutter, F. (2016) · arXiv · doi:10.48550/arxiv.1608.03983 · Cited by 5,309 · cited 1×
“The learning rate is linearly ramped up to 0.01 during the first 5 epochs and then decayed with a cosine schedule [63].”
BackgroundInfluentialInception-v4, Inception-ResNet and the Impact of Residual Connections on Learning
Szegedy, C., Ioffe, S., Vanhoucke, V., & Alemi, A. (2017) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v31i1.11231 · Cited by 4,636 · cited 1×
“Similar to [45], the Steam module is the initial set of operations used to map low-level visual embedding before executing to the SMS module, written as: where Θ represents the learnable parameter matrices.”
MethodSlowFast Networks for Video Recognition
Feichtenhofer, C., Fan, H., Malik, J., & He, K. (2019) · 2019 IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2019.00630 · Cited by 4,007 · cited 1×
“For each block, inspired by [11], [47], we construct multi-branch structures to tackle inputs with different temporal resolutions and feature dimensions accordingly: fast ( R ), moderate ( R T × C ), and slow ( R ), to perceive different motion variations, i.e., the high temporal resolution…”
MethodTemporal Segment Networks: Towards Good Practices for Deep Action Recognition
Wang, L., Xiong, Y., Wang, Z., Qiao, Y., Lin, D., Tang, X., & Van Gool, L. (2016) · Lecture Notes in Computer Science · doi:10.1007/978-3-319-46484-8_2 · Cited by 3,958
Research works printed but not linked: 1
- [82] O. K¨ op ¨ ukl¨ u, A. Gunduz, N. Kose, and G. Rigoll, 'Real-time hand gesture detection and classification using convolutional neural networks,' in 2019 14th IEEE International Conference on Automatic Face & Gesture Recognition (FG 2019) . IEEE, 2019, pp. 1-8. -
Cited by
Gloss-free Sign Language Translation: Improving from Visual-Language Pretraining
Zhou, B., Chen, Z., Clapés, A., Wan, J., Liang, Y., Escalera, S., Lei, Z., & Zhang, D. (2023) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51070.2023.01908 · Cited by 236
“The goal of ISLR is to translate an isolated sign into a corresponding single sign language word [14, 15, 20, 22], which is somewhat similar to the isolated gesture recognition task [33, 18, 9, 38, 40, 39, 36].”
BackgroundSelf-citationRobust Multimodal Learning with Missing Modalities via Parameter-Efficient Adaptation
Reza, M. K., Prater-Bennette, A., & Asif, M. S. (2025) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2024.3476487 · Cited by 78
Zhang, Y., & Wang, Y. (2025) · EURASIP Journal on Image and Video Processing · doi:10.1186/s13640-025-00677-0 · Cited by 25
Li, Y., Chen, H., Feng, G., & Miao, Q. (2023) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51070.2023.01917 · Cited by 24
“Although great progress has been made in this field, the attention has always been drawn to boost performance via better network structure [50, 48, 55, 56] or by introducing extra modalities of data [6, 18, 23, 19].”
BackgroundClassroom Behavior Detection Method Based on PLA-YOLO11n
Zhang, H., Zhou, G., He, W., & Deng, H. (2025) · Sensors · doi:10.3390/s25175386 · Cited by 14
“…multi-scale receptive fields; and (3) the integration of attention mechanisms to dynamically focus on key regions (e.g., hand-raising, head-down, interactive gestures) while suppressing background noise, thereby enhancing the accuracy and stability of action detection and classification [7–9].”
HSPFormer: Hierarchical Spatial Perception Transformer for Semantic Segmentation
Chen, S., Han, T., Zhang, C., Su, J., Wang, R., Chen, Y., Wang, Z., & Cai, G. (2025) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2025.3525542 · Cited by 18
“Currently, many studies have demonstrated that multi-modality data fusion is able to significantly improve the accuracy of semantic segmentation [42], [43], [44], [45].”
Zou, S., Wang, X., Tao, Y., Zeng, K., Li, G., & Xie, X. (2025) · The Journal of the Acoustical Society of America · doi:10.1121/10.0034831 · Cited by 11
RVLF: A Reinforcing Vision-Language Framework for Gloss-Free Sign Language TranslationPreprint
Rao, Z., Zhou, Y., Zhou, B., Huang, Y., Escalera, S., & Wan, J. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2512.07273 · Cited by 6
Self-citationShen, X., Du, H., Sheng, H., Li, L., & Zhang, K. (2025) · International World Wide Web Conference (WWW) · doi:10.1145/3696410.3714525 · Cited by 8
MDSI: Pluggable Multi-strategy Decoupling with Semantic Integration for RGB-D Gesture Recognition
Fang, F., Liao, Z., Kan, Z., Wang, G., & Yang, W. (2025) · Pattern Recognition · doi:10.1016/j.patcog.2025.111653 · Cited by 4
Cross-View Isolated Sign Language Recognition Challenge: Design, Results and Future Research
Shen, X., Du, H., Xu, M., Liu, M., & Yu, X. (2025) · International World Wide Web Conference (WWW) · doi:10.1145/3701716.3717522 · Cited by 4
Linguistic-Driven Partial Semantic Relevance Learning for Skeleton-Based Action Recognition
Chen, Q., Liu, Y., Huang, P., & Huang, J. (2024) · Sensors · doi:10.3390/s24154860 · Cited by 5
Nguyen, T. T. D., Do, T. T. N., Huy, H. Q., Nguyen, P. L., & Vu, T. A. (2025) · IEEE Access · doi:10.1109/access.2025.3601235 · Cited by 3
“Impressively, both UniFormer-S and UsimKD excel on the STU and TED subsets, outperforming complex multi-stream models like KVNet-V [1], UMDR [73], and SlowFast [71], despite being single-stream architectures.”
Dai, J., Feng, Y., Zhu, Y., & Huo, M. (2025) · Signal Image and Video Processing · doi:10.1007/s11760-025-04966-z · Cited by 3
“For RGB-D bimodal methods (e.g., UMDR [20]), color and depth information are integrated.”
SIGMA-ASL: Sensor-Integrated Multimodal Dataset for Sign Language Recognition
Xiao, X., Li, G., Zhao, G., Lin, Q., Ma, W., Wen, H., Wang, Y., & Shen, Y. (2026) · Proceedings of the ACM on Interactive Mobile Wearable and Ubiquitous Technologies · doi:10.1145/3810202 · Cited by 3
Hand Gesture Recognition From an Open-Set Perspective
Zhou, J., Xu, C. Y., & Cheng, L. (2025) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2025.3535363 · Cited by 3
Integrated multi-local and global dynamic perception structure for sign language recognition
Liang, S., Li, Y., Shi, Y., Chen, H., & Miao, Q. (2025) · Pattern Analysis and Applications · doi:10.1007/s10044-024-01403-8 · Cited by 3
“[26 ] proposed the cross-modal Complement Feature Catcher to unearth potential common features within multi-modal information as an auxiliary fusion stream, aiming to improve the output of the subsequent fusion process.”
Synthetic View Augmentation for Sign Language Recognition
Peng, Y., Jiao, P., Zou, H., Min, Y., & Chen, X. (2025) · International World Wide Web Conference (WWW) · doi:10.1145/3701716.3717520 · Cited by 3
“The combination of RGB-D data also outperforms UMDR [37] (57.97% vs. 22.82%), which adopts multi-modal Ablation on multi-modality fusion.”
InfluentialA spatiotemporal dynamic fusion network for surgical action recognition
Bian, G., Peng, Y., Li, Z., Ye, Q., & Ma, R. (2025) · Neurocomputing · doi:10.1016/j.neucom.2025.130808 · Cited by 2
SCOPE: Preserving Modality-Specific Cues to Mitigate Modality Laziness in Multimodal Learning
Yang, J., Zhang, R., Liang, H., & Yuan, K. (2026) · Findings of the Association for Computational Linguistics: ACL (Findings of ACL) · doi:10.18653/v1/2026.findings-acl.1453 · Cited by 2
“Existing methods often achieve this either by imposing distance-based or similarity-based regularization in the embedding space (Li et al., 2023; Zhou et al., 2023a; Wang et al., 2025), or by regulating cross-modal statistical dependence, where mutual-information-related objectives are widely used…”
CM-ASAP: Cross-Modality Adaptive Sensing and Perception for Efficient Hand Gesture Recognition
Hor, S., El‐Khamy, M., Zhou, Y., Arbabian, A., & Lim, S. (2024) · IEEE International Conference on Multimedia Information Processing and Retrieval (MIPR) · doi:10.1109/mipr62202.2024.00039 · Cited by 2
“The recent SOTA [18] proposes a unified multimodal de-coupling and multi-stage re-coupling framework for video representation learning.”
BackgroundMGR-Dark: A Large Multimodal Video Dataset and RGB-IR Benchmark for Gesture Recognition in Darkness
Shi, Y., Li, Y., Liang, S., Chen, H., & Miao, Q. (2024) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3664647.3681267 · Cited by 2
“Numerous datasets[3, 15, 16, 19, 23, 34] and research [4, 17, 24, 33, 36–38] have been dedicated to gesture recognition.”
Learned Sensor Fusion For Robust Human Activity Recognition in Challenging Environments
Conway, M., Reily, B., & Reardon, C. (2024) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros58592.2024.10802830 · Cited by 2
ConTrans-Detect: A Multi-Scale Convolution-Transformer Network for DeepFake Video Detection
Sun, W., Ma, Y., Zhang, H., & Wang, R. (2023) · International Conference on Mechatronics and Machine Vision in Practice (M2VIP) · doi:10.1109/m2vip58386.2023.10413387 · Cited by 2
“Inspired by [36]–[38], our multi-branch Transformer module perceives diverse motion variations, which uses the high temporal dimensions and low channel size as a trade-off.”
BackgroundEnhancing Skeleton-Based Action Recognition with Feature Maps from Pose Estimation Networks
Hao, W. E. N., Zhe-Ming, L. U., Shen, F., Lu, Z., Zheng, Y., & Jialin, C. U. I. (2025) · IEICE Transactions on Fundamentals of Electronics Communications and Computer Sciences · doi:10.1587/transfun.2024eap1162 · Cited by 2
Hu, Y., Wang, S., Hu, Q., & Li, J. (2025) · IEEE Sensors Journal · doi:10.1109/jsen.2025.3576794 · Cited by 1
“Some studies have explored the potential of transformer structures to enhance gesture recognition [17], [36].”
CISampler: Correlated Information Guided Frame Sampling for Gesture Recognition in Video
Shi, Y., Li, Y., Chen, H., Liang, S., & Miao, Q. (2024) · ACM International Conference on Multimedia in Asia (MMAsia) · doi:10.1145/3696409.3700224 · Cited by 1
“Recently, some researchers have explored the benefits of transformers[3] in gesture recognition, applying network architecture transformers[19, 51, 52] to enhance recognition performance.”
InfluentialCross-View Isolated Sign Language Recognition with Graph
Wang, X. (2025) · International World Wide Web Conference (WWW) · doi:10.1145/3701716.3717519 · Cited by 1
“Our method, ISLR-Graph, achieves an average Top-1 accuracy of 33.97%, significantly outperforming the baseline model UMDR [26] (22.82% for Pixel + Depth).”
InfluentialMultimodal Disentanglement and Distillation Representation for Gesture RecognitionPreprint
Shi, Y., Li, Y., Chen, H., Liang, S., & Miao, Q. (2024) · SSRN Electronic Journal · doi:10.2139/ssrn.4963605
Vision-Based Abnormal Action Dataset for Recognising Body Motion Disorders
Ying, J., Shen, X., & Yu, X. (2024) · Lecture notes in computer science · doi:10.1007/978-981-96-0351-0_33
Feghoul, K. (2024) · theses.fr (ABES) · doi:10.70675/bfc0d4bazb8eaz4305z957bze99ec2b0e6f8
LAGD-Net: Second-Order Motion Representation Learning for RGB-D Motion Recognition
Zhou, B., Huang, Y., Liang, Y., Zhang, D., & Wan, J. (2024) · International Conference on Signal Processing · doi:10.1109/icsp62129.2024.10846675
“Zhou et al. [7] introduced a framework that decouples and then recouples multi-modal representations across stages to better understand space-time correlations.”
InfluentialSelf-citationAction recognition from egocentric videos using RGB + Depth modalities
Vishwakarma, P. K., Singh, D. K., & Sahu, A. (2025) · International Conference on Recent Advances in Information Technology · doi:10.1109/rait65068.2025.11089208
“We compare our method with seven recent approaches, including [13], [32]–[35], [37], on the eight (one common [32]) MECCANO dataset sequences.”
CAMF: A Cross-modal Attention and Multi-stage Fusion Network for Multi-modal Vessel Identification
Wang, S., Chen, J., Lin, Y., & Xu, C. (2025) · Lecture notes of the Institute for Computer Sciences, Social Informatics and Telecommunications Engineering · doi:10.1007/978-3-032-00274-7_16
Multimodal Transfer Learning for Privacy in Human Activity Recognition
Rolfsjord, S., Fatima, S., von Arnim, H. A., & Baselizadeh, A. (2025) · IEEE RO-MAN · doi:10.1109/ro-man63969.2025.11217600 · Cited by 1
“For RGB and depth, we use the near state-of-the-art backbones from UMDR [24], which are trained on the NTU RGB+D 120 dataset.”
Action Recognition in RGB-D Videos via an End-to-End Cross-Modal Attention Transformer
Ma, Y., Zhou, B., Zhang, H., Zhang, X., He, X., & Wang, R. (2026) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2026.3674923
“V IDEO action recognition serves as the cornerstone of video understanding, finding applications in diverse fields such as human-robot interaction, healthcare, and video surveillance systems [1], [2], [3].”
InfluentialSelf-citationAn Open-Set Framework for Robust Hand Gesture Classification
Venkateswarlu, B., Susmitha, M., Varalakshmi, R., Muneer, S. K., & Pavan, T. (2026) · Iconic Research and Engineering Journals · doi:10.64388/irev9i9-1715462
DEFT-DPT: A Lightweight Multimodal Framework for Egocentric Video Action Recognition
Vishwakarma, P. K., Singh, D. K., & Sahu, A. (2026) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2026.3707206
“However, most multimodal frameworks [10], [25], [26] still use static or early-fusion schemes that fail to capture the changing importance of frames and modalities during egocentric interactions.”
LGMFN: A local–global multimodal fusion network for robust and efficient human action recognition
Li, L., Hu, Z., Ji, G., Qi, X., & Xiao, J. Y. (2026) · Journal of Visual Communication and Image Representation · doi:10.1016/j.jvcir.2026.104962
RVLF: A Reinforcing Vision-Language Framework for Gloss-Free Sign Language Translation
Rao, Z., Zhou, Y., Zhou, B., Huang, Y., Escalera, S., & Wan, J. (2026) · doi:10.1109/cvprf71909.2026.00886
Self-citationVo, D. T. D., Truong, N.-N., Vy, D. N.-L., Nguyen, V. D., & Le, M.-H. (2026) · International Conference on Green Technology and Sustainable Development (GTSD) · doi:10.1109/gtsd70918.2026.11722506
“Notably, while UMDR [39] achieves a strong 95.00% via an auxiliary late-fusion ensemble, U2M-Net elevates this to 96.00%.”
InfluentialIbrahim, D. M., & El-Ghaish, H. (2026) · Journal of King Saud University - Computer and Information Sciences · doi:10.1007/s44443-026-01310-4
Metadata
- Authors
- Zhou, B., Wang, P., Wan, J., Liang, Y., & Wang, F.
- Affiliations
- Macau University of Science and Technology (MO)
- Amazon (United States) (US)
- Chinese Academy of Sciences (CN)
- Alibaba Group (United States) (US)
- Venue
- IEEE Transactions on Pattern Analysis and Machine Intelligence
- Published
- 2023-10
- Volume
- 45
- Issue
- 10
- Pages
- 11428-11442
- ISSN
- 0162-8828 (ISSN-L)
- 2160-9292
- 1939-3539
- OpenAlex
- W4376132550
- Cited by
- 38
- References
- 86
- Citations by year
- 2026: 7 · 2025: 19 · 2024: 9 · 2023: 3
- FWCI
- 3.45
- Type
- journal article
- Language
- en
- Open access
- green
- arxiv.org
- Licence
- Locations
- IEEE Transactions on Pattern Analysis and Machine Intelligence (publishedVersion)
- PubMed (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- arXiv (Cornell University)
- Funders
- National Key Research and Development Plan (2021YFE0205700)
- External cooperation key project of Chinese Academy Sciences (173211KYSB20200002)
- Science and Technology Development Fund of Macau (0123/2022/A3, 0070/2020/AMJ, 0004/2020/A1)
- Guangdong Provincial Key R&D Programme (2019B010148001)
- Open Research Projects of Zhejiang Lab (2021KH0AB07)
- CCF-Zhipu AI Large Model OF (202219)
- Alibaba Group
- Topics
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Human Pose and Action Recognition
- Physical Sciences › Computer Science › Human-Computer Interaction › Hand Gesture Recognition Systems
- Physical Sciences › Computer Science › Computer Graphics and Computer-Aided Design › Human Motion and Animation
- Keywords
- multimodal fusion, mixup, spatio-temporal representation learning, decision fusion
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.