Spatial Temporal Graph Convolutional Networks for Skeleton-Based Action Recognition
Open accessAbstract
Dynamics of human body skeletons convey significant information for human action recognition. Conventional approaches for modeling skeletons usually rely on hand-crafted parts or traversal rules, thus resulting in limited expressive power and difficulties of generalization. In this work, we propose a novel model of dynamic skeletons called Spatial-Temporal Graph Convolutional Networks (ST-GCN), which moves beyond the limitations of previous methods by automatically learning both the spatial and temporal patterns from data. This formulation not only leads to greater expressive power but also stronger generalization capability. On two large datasets, Kinetics and NTU-RGBD, it achieves substantial improvements over mainstream methods.
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Cites
ImageNet classification with deep convolutional neural networks
Krizhevsky, A., Sutskever, I., & Hinton, G. E. (2017) · Communications of the ACM · doi:10.1145/3065386 · Cited by 112,155 · cited 1× · see citing works
“In tasks such as image object recognition, the hierarchical representation and locality are usually achieved by the intrinsic properties of convolutional neural networks (Krizhevsky, Sutskever, and Hinton 2012), rather than manually assigning object parts.”
MethodSemi-supervised classification with graph convolutional networks
Kipf, T. N., & Welling, M. (2016) · arXiv · doi:10.48550/arxiv.1609.02907 · Cited by 24,853 · cited 8× · see citing works
“In the single frame case, ST-GCN with the first partitioning strategy can be implemented with the following formula (Kipf and Welling 2017) where Here the weight vectors of multiple output channels are stacked to form the weight matrix W .”
BackgroundMethodInfluentialLearning Spatiotemporal Features with 3D Convolutional Networks
Tran, D., Bourdev, L., Fergus, R., Torresani, L., & Paluri, M. (2015) · 2015 IEEE International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2015.510 · Cited by 9,593 · cited 2× · see citing works
“In general, human action can be recognized from multiple modalities(Simonyan and Zisserman 2014; Tran et al. 2015; Wang, Qiao, and Tang 2015; Wang et al. 2016; Zhao et al. 2017), such as appearance, depth, optical flows, and body skeletons (Du, Wang, and Wang 2015; Liu et al.”
BackgroundRealtime Multi-person 2D Pose Estimation Using Part Affinity Fields
Cao, Z., Simon, T., Wei, S.-E., & Sheikh, Y. (2017) · 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2017.143 · Cited by 7,539 · cited 3× · see citing works
“Skeleton and joint trajectories of human bodies are robust to illumination change and scene variation, and they are easy to obtain owing to the highly accurate depth sensors or pose estimation algorithms (Shotton et al. 2011; Cao et al. 2017a).”
BackgroundMethodDeformable Convolutional Networks
Dai, J., Qi, H., Xiong, Y., Li, Y., Zhang, G., Hu, H., & Wei, Y. (2017) · 2017 IEEE International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2017.89 · Cited by 7,513 · cited 1× · see citing works
“More detailed explanation and other applications of this formulation can be found in (Dai et al. 2017).”
BackgroundSelf-citationAdvances In Deep Learning On Graphs (Gsp'18 Workshop)Preprint
Defferrard, M., Bresson, X., & Vandergheynst, P. (2016) · arXiv (Cornell University) · doi:10.48550/arxiv.1606.09375 · Cited by 6,912 · cited 2× · see citing works
“…graphs of arbitrary structures, have received increasing attention and successfully been adopted in a number of applications, such as image classification (Bruna et al. 2014), document classifi-cation (Defferrard, Bresson, and Vandergheynst 2016), and semi-supervised learning (Kipf and Welling 2017).”
BackgroundTwo-Stream Convolutional Networks for Action Recognition in Videos
Simonyan, K., & Zisserman, A. (2014) · arXiv · doi:10.48550/arxiv.1406.2199 · Cited by 6,756 · cited 1× · see citing works
“In general, human action can be recognized from multiple modalities(Simonyan and Zisserman 2014; Tran et al. 2015; Wang, Qiao, and Tang 2015; Wang et al. 2016; Zhao et al. 2017), such as appearance, depth, optical flows, and body skeletons (Du, Wang, and Wang 2015; Liu et al. 2016).”
BackgroundTemporal Segment Networks: Towards Good Practices for Deep Action Recognition
Wang, L., Xiong, Y., Wang, Z., Qiao, Y., Lin, D., Tang, X., & Gool, L. V. (2016) · Lecture notes in computer science · doi:10.1007/978-3-319-46484-8_2 · Cited by 4,164 · cited 4× · see citing works
“We train the standard TSN (Wang et al. 2016) models from scratches on Kinetics with RGB and optical flow models.”
BackgroundMethodInfluentialSelf-citationSpectral Networks and Locally Connected Networks on Graphs
Bruna, J., Zaremba, W., Szlam, A., & LeCun, Y. (2013) · arXiv · doi:10.48550/arxiv.1312.6203 · Cited by 3,983 · see citing works
Gated Graph Sequence Neural NetworksPreprint
Li, Y., Tarlow, D., Brockschmidt, M., & Zemel, R. (2015) · International Conference on Learning Representations · doi:10.48550/arxiv.1511.05493 · Cited by 3,735 · cited 4× · see citing works
“The discussed neural network architectures include both recurrent neural networks (Tai, Socher, and Manning 2015; Van Oord, Kalchbrenner, and Kavukcuoglu 2016) and convolutional neural networks (CNNs) (Bruna et al. 2014; Henaff, Bruna, and LeCun 2015; Duvenaud et al. 2015; Li et al. 2016; Defferrard, Bresson, and Vand…”
BackgroundInfluentialThe Kinetics Human Action Video DatasetPreprint
Kay, W., Carreira, J., Simonyan, K., Zhang, B., Hillier, C., Vijayanarasimhan, S., Viola, F., Green, T., Back, T., Natsev, P., et al. (2017) · arXiv (Cornell University) · doi:10.48550/arxiv.1705.06950 · Cited by 3,711 · cited 13× · see citing works
“Weevaluate the recognition performance by top-1 and top-5 classification accuracy as recommended by the dataset authors (Kay et al. 2017).”
BackgroundMethodInfluentialNTU RGB+D: A Large Scale Dataset for 3D Human Activity Analysis
Shahroudy, A., Liu, J., Ng, T.-T., & Wang, G. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.115 · Cited by 3,148 · cited 9× · see citing works
“All rights reserved. attempt to leverage the natural connections between joints have been developed (Shahroudy et al. 2016; Du, Wang, and Wang 2015).”
BackgroundMethodInfluentialPixel Recurrent Neural NetworksPreprint
van den Oord, A., Kalchbrenner, N., & Kavukcuoglu, K. (2016) · International Conference on Machine Learning · doi:10.48550/arxiv.1601.06759 · Cited by 3,111 · see citing works
Improved Semantic Representations From Tree-Structured Long Short-Term Memory Networks
Tai, K. S., Socher, R., & Manning, C. D. (2015) · Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Processing (Volume 1: Long Papers) · doi:10.3115/v1/p15-1150 · Cited by 3,081 · see citing works
Convolutional Networks on Graphs for Learning Molecular Fingerprints
Duvenaud, D., Maclaurin, D., Aguilera-Iparraguirre, J., Gómez-Bombarelli, R., Hirzel, T., Aspuru-Guzik, A., & Adams, R. P. (2015) · arXiv · doi:10.48550/arxiv.1509.09292 · Cited by 2,628 · cited 2× · see citing works
“…spectral perspective , where the locality of the graph convolution is considered in the form of spectral analysis (Henaff, Bruna, and Le-Cun 2015; Duvenaud et al. 2015; Li et al. 2016; Kipf and Welling 2017); 2) the spatial perspective , where the convolution filters are applied directly on the…”
BackgroundLearning Convolutional Neural Networks for GraphsPreprint
Niepert, M., Ahmed, M., & Kutzkov, K. (2016) · International Conference on Machine Learning · doi:10.48550/arxiv.1605.05273 · Cited by 2,311 · cited 2× · see citing works
“The solution to this problem is first investigated in (Niepert, Ahmed, and Kutzkov 2016), where the order is defined by a graph labeling process in the neighbor graph around the root node.”
BackgroundMethodReal-time human pose recognition in parts from single depth images
Shotton, J., Sharp, T., Kipman, A., Fitzgibbon, A. W., Finocchio, M., Blake, A., Cook, M., & Moore, R. (2013) · Communications of the ACM · doi:10.1145/2398356.2398381 · Cited by 2,050 · see citing works
Hierarchical recurrent neural network for skeleton based action recognition
Du, Y., Wang, W., & Wang, L. (2015) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr.2015.7298714 · Cited by 1,962 · cited 2× · see citing works
“All rights reserved. attempt to leverage the natural connections between joints have been developed (Shahroudy et al. 2016; Du, Wang, and Wang 2015).”
BackgroundHuman Action Recognition by Representing 3D Skeletons as Points in a Lie Group
Vemulapalli, R., Arrate, F., & Chellappa, R. (2014) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr.2014.82 · Cited by 1,506 · see citing works
Spatio-Temporal LSTM with Trust Gates for 3D Human Action Recognition
Liu, J., Shahroudy, A., Xu, D., & Wang, G. (2016) · Lecture notes in computer science · doi:10.1007/978-3-319-46487-9_50 · Cited by 1,353 · cited 3× · see citing works
“In general, human action can be recognized from multiple modalities(Simonyan and Zisserman 2014; Tran et al. 2015; Wang, Qiao, and Tang 2015; Wang et al. 2016; Zhao et al. 2017), such as appearance, depth, optical flows, and body skeletons (Du, Wang, and Wang 2015; Liu et al. 2016).”
BackgroundMethodDeep Convolutional Networks on Graph-Structured Data
Henaff, M., Bruna, J., & LeCun, Y. (2015) · arXiv · doi:10.48550/arxiv.1506.05163 · Cited by 1,255 · see citing works
Action recognition with trajectory-pooled deep-convolutional descriptors
Wang, L., Qiao, Y., & Tang, X. (2015) · 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2015.7299059 · Cited by 1,194 · see citing works
Zhu, W., Lan, C., Xing, J., Zeng, W.-J., Li, Y.-H., Shen, L., & Xie, X.-H. (2016) · AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v30i1.10451 · Cited by 934 · cited 2× · see citing works
“These works have been using recurrent neural networks (Shahroudy et al. 2016; Zhu et al. 2016; Liu et al. 2016; Zhang, Liu, and Xiao 2017) and temporal CNNs (Li et al.”
BackgroundA New Representation of Skeleton Sequences for 3D Action Recognition
Ke, Q., Bennamoun, M., An, S., Sohel, F., & Boussaid, F. (2017) · 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2017.486 · Cited by 793 · cited 10× · see citing works
“Our ST-GCN is the first to apply graph CNNs to the task of skeleton based action recognition.”
BackgroundMethodInfluentialInterpretable 3D Human Action Analysis with Temporal Convolutional Networks
Kim, T. S., & Reiter, A. (2017) · IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) · doi:10.1109/cvprw.2017.207 · Cited by 696 · see citing works
Modeling video evolution for action recognition
Fernando, B., Gavves, E., Jose Oramas, M., Ghodrati, A., & Tuytelaars, T. (2015) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr.2015.7299176 · Cited by 573 · cited 2× · see citing works
“The first is the feature encoding approach on hand-crafted features (Fernando et al. 2015), referred to as “Feature Encoding” in Table 2.”
MethodDifferential Recurrent Neural Networks for Action Recognition
Veeriah, V., Zhuang, N., & Qi, G.-J. (2015) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2015.460 · Cited by 525 · see citing works
Skeleton-based action recognition with convolutional neural networks
Li, C., Zhong, Q., Xie, D., & Pu, S. (2017) · IEEE International Conference on Multimedia & Expo Workshops (ICMEW) · doi:10.1109/icmew.2017.8026285 · Cited by 392 · cited 8× · see citing works
“Our ST-GCN is the first to apply graph CNNs to the task of skeleton based action recognition.”
BackgroundInfluentialOn Geometric Features for Skeleton-Based Action Recognition Using Multilayer LSTM Networks
Zhang, S., Liu, X., & Xiao, J. (2017) · IEEE Winter Conference on Applications of Computer Vision · doi:10.1109/wacv.2017.24 · Cited by 318 · see citing works
Temporal Action Detection with Structured Segment Networks
Zhao, Y., Xiong, Y., Wang, L., Wu, Z., Tang, X., & Lin, D. (2019) · International Journal of Computer Vision · doi:10.1007/s11263-019-01211-2 · Cited by 90 · cited 2× · see citing works
“In general, human action can be recognized from multiple modalities(Simonyan and Zisserman 2014; Tran et al. 2015; Wang, Qiao, and Tang 2015; Wang et al. 2016; Zhao et al. 2017), such as appearance, depth, optical flows, and body skeletons (Du, Wang, and Wang 2015; Liu et al.”
BackgroundSelf-citationChen, T., Wang, Z., Li, L., Li, D., Li, Z., Chang, X., Bian, P., Liang, G., Wang, Q., & Xie, T. (2025) · Schloss Dagstuhl – Leibniz-Zentrum für Informatik · doi:10.4230/lipics.ecoop.2025.6
Flagged first, then by citations
1 of the 32 references couldn’t be matched to a known paper.
Research works printed but not linked: 4
- [Cao et al. 2017b] Cao, Z.; Simon, T.; Wei, S.-E.; and Sheikh, Y. 2017b. Realtime multi-person 2d pose estimation using part affinity fields. In CVPR .
- [Defferrard, Bresson, and Vandergheynst 2016] Defferrard, M.; Bresson, X.; and Vandergheynst, P. 2016. Convolutional neural networks on graphs with fast localized spectral filtering. In NIPS .
- [Hussein et al. 2013] Hussein, M. E.; Torki, M.; Gowayyed, M. A.; and El-Saban, M. 2013. Human action recognition using a temporal hierarchy of covariance descriptors on 3d joint locations. In IJCAI .
- [Wang et al. 2012] Wang, J.; Liu, Z.; Wu, Y.; and Yuan, J. 2012. Mining actionlet ensemble for action recognition with depth cameras. In CVPR . IEEE.
Cited by
Sorted by citations
Graph neural networks: A review of methods and applications
Zhou, J., Cui, G., Hu, S., Zhang, Z., Yang, C., Liu, Z., Wang, L., Li, C., & Sun, M. (2020) · AI Open · doi:10.1016/j.aiopen.2021.01.001 · Cited by 6,880 · see citing works
Connecting the Dots: Multivariate Time Series Forecasting with Graph Neural Networks
Wu, Z.-H., Pan, S., Long, G., Jiang, J., Chang, X.-J., & Zhang, C.-Q. (2020) · Knowledge Discovery and Data Mining · doi:10.1145/3394486.3403118 · Cited by 2,447 · see citing works
“For spatial-temporal graph neural networks, spatial dependencies among nodes are captured by graph convolutions, while temporal dependencies among historical states are preserved by recurrent neural networks [13, 17] or 1D convolutions [22, 23].”
BackgroundGraph convolutional networks: a comprehensive review
Zhang, S., Tong, H., Xu, J., & Maciejewski, R. (2019) · Computational Social Networks · doi:10.1186/s40649-019-0069-y · Cited by 1,888 · see citing works
“In [81], a spatial-temporal graph convolutional model is designed to eliminate the need of hand-crafted part assignment and can achieve a greater expressive power.”
BackgroundLearning to Simulate Complex Physics with Graph NetworksPreprint
Sanchez-Gonzalez, A., Godwin, J., Pfaff, T., Ying, R., Leskovec, J., & Battaglia, P. (2020) · International Conference on Machine Learning · doi:10.48550/arxiv.2002.09405 · Cited by 1,739 · see citing works
“…et al., 2019; Sanchez-Gonzalez et al., 2019), as well as non-physical systems, such as multi-agent dynamics (Tacchetti et al., 2018; Sun et al., 2019), algorithm execution (Velikovi et al., 2020), and other dynamic graph settings (Trivedi et al., 2019; 2017; Yan et al., 2018; Manessi et al., 2020).”
BackgroundDisentangling and Unifying Graph Convolutions for Skeleton-Based Action Recognition
Liu, Z., Zhang, H., Chen, Z., Wang, Z., & Ouyang, W. (2020) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr42600.2020.00022 · Cited by 1,307 · see citing works
“In particular, skeleton-based human action recognition involves predicting actions from skeleton representations of human bodies instead of raw RGB videos, and the significant results seen in recent work [50, 33, 32, 34, 21, 20, 54, 35] have proven its merits.”
BackgroundMethodResultInfluentialMohamed, A. A., Qian, K., Elhoseiny, M., & Claudel, C. (2020) · Computer Vision and Pattern Recognition · doi:10.1109/cvpr42600.2020.01443 · Cited by 1,100 · see citing works
“One of the development related to our work is the ST-GCNN [27].”
BackgroundDeep Learning-based Human Pose Estimation: A Survey
Zheng, C., Wu, W., Chen, C., Yang, T., Zhu, S., Shen, J., Kehtarnavaz, N., & Shah, M. A. (2023) · ACM Computing Surveys · doi:10.1145/3603618 · Cited by 836 · see citing works
“[275] leveraged the dynamic skeleton modality of pose for action recognition.”
BackgroundChannel-wise Topology Refinement Graph Convolution for Skeleton-Based Action Recognition
Chen, Y.-X., Zhang, Z.-Q., Yuan, C.-F., Li, B., Deng, Y., & Hu, W.-M. (2021) · IEEE International Conference on Computer Vision · doi:10.1109/iccv48922.2021.01311 · Cited by 1,057 · see citing works
“[31] proposed an ST-GCN which predefines topology according to human body structure and the topology is fixed in both training and testing phase.”
BackgroundMethodInfluentialSkeleton-Based Action Recognition With Shift Graph Convolutional Network
Cheng, K., Zhang, Y., He, X., Chen, W., Cheng, J., & Lu, H. (2020) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr42600.2020.00026 · Cited by 1,046 · see citing works
“For example, ST-GCN [34] costs 16.”
BackgroundMethodInfluentialUnbiased Scene Graph Generation From Biased Training
Tang, K., Niu, Y., Huang, J., Shi, J., & Zhang, H. (2020) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr42600.2020.00377 · Cited by 784 · see citing works
“— yet on which are our core efforts made [69, 53, 6], then pretend that there is a graph — nothing but a sparse object layout with binary links, and finally shroud it into graph neural networks [63] for merely more contextual object representations [65, 17, 54].”
Human Action Recognition From Various Data Modalities: A Review
Sun, Z., Ke, Q., Rahmani, H., Bennamoun, M., Wang, G., & Liu, J. (2022) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2022.3183112 · Cited by 691 · see citing works
“Thus many recent works on skeleton-based HAR used skeleton data obtained from depth maps [151] or RGB videos [152].”
BackgroundMethodInfluentialRevisiting Skeleton-based Action Recognition
Duan, H., Zhao, Y., Chen, K., Lin, D., & Dai, B. (2022) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52688.2022.00298 · Cited by 902 · see citing works
“Kinetics-400.”
BackgroundMethodInfluentialSelf-citationZhang, J., & Tao, D. (2021) · IEEE Internet of Things Journal · doi:10.1109/jiot.2020.3039359 · Cited by 774 · see citing works
“To model the long-term temporal dependencies and dynamics, as well as spatial structures within the skeleton sequence, different neural networks have been exploited for action recognition, such as the deep recurrent neural network (RNN) [116], CNN [117], and deep graph convolutional networks (GCNs) [118].”
MethodLearning Dynamics and Heterogeneity of Spatial-Temporal Graph Data for Traffic Forecasting
Guo, S., Lin, Y., Wan, H., Li, X., & Cong, G. (2021) · IEEE Transactions on Knowledge and Data Engineering · doi:10.1109/tkde.2021.3056502 · Cited by 774 · see citing works
“Depending on the choice of forward computation function, the STGNNs-based meth-ods can be divided into two categories, RNNs-based approaches [6], [7], [8] and CNNs-based approaches [9], [10], [11], [12].”
MethodInfluentialAction-Conditioned 3D Human Motion Synthesis with Transformer VAE
Petrovich, M., Black, M. J., & Varol, G. (2021) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv48922.2021.01080 · Cited by 506 · see citing works
Vision GNN: An Image is Worth Graph of NodesPreprint
Han, K., Wang, Y.-H., Guo, J.-Y., Tang, Y.-H., & Wu, E. (2022) · Neural Information Processing Systems · doi:10.48550/arxiv.2206.00272 · Cited by 689 · see citing works
“By processing the naturally formed graph of linked human joints, GCN was utilized on human action recognition task [23, 62].”
MethodSTGAT: Modeling Spatial-Temporal Interactions for Human Trajectory Prediction
Huang, Y., Bi, H., Li, Z., Mao, T., & Wang, Z. (2019) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2019.00637 · Cited by 697 · see citing works
“Recently, some meth-ods [27] [40] in the field of action recognition have made significant progress by applying GNNs to spatial-temporal”
BackgroundTraffic Flow Prediction via Spatial Temporal Graph Neural Network
Wang, X., Ma, Y., Wang, Y., Jin, W., Wang, X., Tang, J., Jia, C., & Yu, J. (2020) · Proceedings of The Web Conference 2020 · doi:10.1145/3366423.3380186 · Cited by 644 · see citing works
Not only Look, But Also Listen: Learning Multimodal Violence Detection Under Weak Supervision
Wu, P., Liu, J., Shi, Y., Sun, Y., Shao, F., Wu, Z., & Yang, Z. (2020) · Lecture notes in computer science · doi:10.1007/978-3-030-58577-8_20 · Cited by 521 · see citing works
“For instance, GCN are used for temporal action localization [50], video classification [37,41], anomaly detection [51], skeleton-based action recognition [33,45], point cloud semantic segmentation [21], image captioning [46], and so on.”
MethodA Systematic Review on Affective Computing: Emotion Models, Databases, and Recent Advances
Wang, Y., Song, W., Tao, W., Liotta, A., Yang, D., Li, X., Gao, S., Sun, Y., Ge, W., Zhang, W., et al. (2022) · Information Fusion · doi:10.48550/arxiv.2203.06935 · Cited by 608
“In the combination of CNNs and RNNs, ConvNet-RNN networks with the cascaded fusion [230,283] or the ensemble strategy [215] have been proposed for FER, by simultaneously capturing textural and temporal information.”
MethodInfluentialLearning Trajectory Dependencies for Human Motion Prediction
Mao, W., Liu, M.-M., Salzmann, M., & Li, H. (2019) · IEEE International Conference on Computer Vision · doi:10.1109/iccv.2019.00958 · Cited by 591 · see citing works
“This strategy has been employed for action recognition [25], by using a GCN to capture the temporal and spatial dependencies of human joints via a graph defined on temporally connected kinematic trees.”
MethodSkeleton-Based Action Recognition With Multi-Stream Adaptive Graph Convolutional Networks
Shi, L., Zhang, Y., Cheng, J., & Lu, H. (2020) · IEEE Transactions on Image Processing · doi:10.1109/tip.2020.3028207 · Cited by 619 · see citing works
Graph Convolutional Networks for Temporal Action Localization
Zeng, R., Huang, W., Gan, C., Tan, M., Rong, Y., Zhao, P., & Huang, J. (2019) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2019.00719 · Cited by 559 · see citing works
“Thanks to its effectiveness, GCNs have been successfully applied to several research areas in computer vision, such as skeleton-based action recognition [47], person re-identification [32], and video classification [45].”
BackgroundMethodXia, F., Sun, K., Yu, S., Aziz, A., Wan, L., Pan, S., & Liu, H. (2021) · IEEE Transactions on Artificial Intelligence · doi:10.1109/tai.2021.3076021 · Cited by 495 · see citing works
“studies in this direction have discussed the use of GCNs, the combination of GCNs with RNN or CNN, and recursive structures for graph structures [130], [131], [159].”
BackgroundMethodExploiting Spatial-Temporal Relationships for 3D Pose Estimation via Graph Convolutional Networks
Cai, Y., Ge, L., Liu, J., Cai, J., Cham, T., Yuan, J., & Magnenat‐Thalmann, N. (2019) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2019.00236 · Cited by 522 · see citing works
“To tackle this issue, inspired from the previous studies [33, 50] that take the convolutional operator with a larger kernel size, we made modifications to the generic graph convolutional operations.”
MethodA Comprehensive Survey of Vision-Based Human Action Recognition Methods
Zhang, H.-B., Zhang, Y.-X., Zhong, B., Lei, Q., Yang, L., Du, J.-X., & Chen, D.-S. (2019) · Sensors · doi:10.3390/s19051005 · Cited by 503 · see citing works
FineGym: A Hierarchical Video Dataset for Fine-Grained Action Understanding
Shao, D., Zhao, Y., Dai, B., & Lin, D. (2020) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr42600.2020.00269 · Cited by 369 · see citing works
“3 ST-GCN [48] Pose 11.”
MethodSelf-citationConstructing Stronger and Faster Baselines for Skeleton-Based Action Recognition
Song, Y., Zhang, Z., Shan, C., & Wang, L. (2022) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2022.3157033 · Cited by 544 · see citing works
Assembly101: A Large-Scale Multi-View Video Dataset for Understanding Procedural Activities
Şener, F., Chatterjee, D., Shelepov, D., He, K., Singhania, D., Wang, R. Y., & Yao, A. (2022) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52688.2022.02042 · Cited by 217 · see citing works
““2s-AGCN [41] w/ context” extends each boundary by 0.5 seconds; the extension improves action accuracy significantly.”
BackgroundInfluentialA Review of Graph Neural Networks and Their Applications in Power Systems
Liao, W., Bak‐Jensen, B., Pillai, J. R., Wang, Y., & Wang, Y. (2022) · Journal of Modern Power Systems and Clean Energy · doi:10.35833/mpce.2021.000058 · Cited by 441 · see citing works
“For example, a PGC layer and a one-dimensional convolutional layer are used to build spatial-temporal block in [55].”
MethodMotionBERT: A Unified Perspective on Learning Human Motion Representations
Zhu, W., Ma, X., Liu, Z., Liu, L., Wu, W., & Wang, Y. (2023) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51070.2023.01385 · Cited by 377 · see citing works
“volutional Networks (ST-GCN) have been applied to modeling spatio-temporal relationship of human joints in both 3D pose estimation [13, 116] and action recognition [96, 123], but their connections have not been fully explored.”
BackgroundMethodInfluentialSkeleton-based action recognition via spatial and temporal transformer networks
Plizzari, C., Cannici, M., & Matteucci, M. (2021) · Computer Vision and Image Understanding · doi:10.1016/j.cviu.2021.103219 · Cited by 409 · see citing works
“Finally, in Figure 5b we also compare the entire stream architectures, i.e., ST-GCN (Yan et al. (2018)) and 1s-AGCN (Shi et al. (2019b)) with the proposed S-TR and T-TR streams in terms of parameters.”
BackgroundMethodInfluentialMulti-Scale Spatial Temporal Graph Convolutional Network for Skeleton-Based Action Recognition
Chen, Z., Li, S., Yang, B., Li, Q., & Liu, H. (2021) · AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v35i2.16197 · Cited by 393 · see citing works
“ST-GCN (Yan, Xiong, and Lin 2018) 31.6 53.7 AS-GCN (Li et al. 2019a) 34.8 56.3 2s AGCN (Shi et al. 2019c) 36.1 58.7 4s DGNN (Shi et al. 2019a) 36.9 56.9 2s NAS-GCN (Peng et al. 2020) 37.1 60.1 2s MS-G3D (Liu et al. 2020) 38.0 60.9 Js”
MethodInfluentialSGCN:Sparse Graph Convolution Network for Pedestrian Trajectory Prediction
Shi, L., Wang, L., Long, C., Zhou, S., Zhou, M., Niu, Z., & Hua, G. (2021) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr46437.2021.00888 · Cited by 378 · see citing works
“To summarize, both Social-STGCNN and SGAN predict overlapping distributions and deviate from the ground-truth, while our predicted distributions exhibit less overlapping and have a better tendency along the ground-truth.”
BackgroundInfluentialDynamic GCN: Context-enriched Topology Learning for Skeleton-based Action Recognition
Ye, F., Pu, S., Zhong, Q., Li, C., Xie, D., & Tang, H.-M. (2020) · ACM Multimedia · doi:10.1145/3394171.3413941 · Cited by 384 · see citing works
“We follow the same train-validation split as [33].”
BackgroundMethodInfluentialDynamic Multiscale Graph Neural Networks for 3D Skeleton Based Human Motion Prediction
Li, M., Chen, S.-H., Zhao, Y., Zhang, Y., Wang, Y.-F., & Tian, Q. (2020) · Computer Vision and Pattern Recognition · doi:10.1109/cvpr42600.2020.00029 · Cited by 382 · see citing works
“Graph deep learning: Graphs, expressing data associated with non-grid structures, preserve the dependencies among internal nodes [46, 40, 39].”
BackgroundMethodInfoGCN: Representation Learning for Human Skeleton-based Action Recognition
Chi, H.-G., Ha, M. H., Chi, S.-G., Lee, S. W., Huang, Q.-X., & Ramani, K. (2022) · Computer Vision and Pattern Recognition · doi:10.1109/cvpr52688.2022.01955 · Cited by 376 · see citing works
“Since the introduction of GCN [21], various approaches exploiting the graph structure of extrinsic topology have been introduced [26, 33, 43, 56].”
BackgroundMethodInfluentialLearning Graph Convolutional Network for Skeleton-based Human Action Recognition by Neural Searching
Peng, W., Hong, X.-P., Chen, H., & Zhao, G.-Y. (2019) · AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v34i03.5652 · Cited by 375 · see citing works
“To keep consistent with the current state-of-the-art GCN methods (Yan, Xiong, and Lin 2018; Shi et al. 2019) 2 , we introduce ten GCN blocks into our network for both searching and training steps.”
MethodResultInfluentialSong, Y., Zhang, Z., Shan, C., & Wang, L. (2020) · ACM Multimedia · doi:10.1145/3394171.3413802 · Cited by 363 · see citing works
“, ST-GCN [31], the PA-ResGCN achieves over 10% and 20% relative performance increases with the similar model size on the two datasets, NTU Frame n Frame n+1”
BackgroundMethodInfluentialH2O: Two Hands Manipulating Objects for First Person Interaction Recognition
Kwon, T., Tekin, B., Stühmer, J., Bogo, F., & Pollefeys, M. (2021) · IEEE International Conference on Computer Vision · doi:10.1109/iccv48922.2021.00998 · Cited by 356 · see citing works
“ST-GCN works on a single graph entity, e.g. human skeleton, and models intra-skeleton connections with a fixed adjacency matrix.”
BackgroundMethodInfluentialWhen physics meets machine learning: a survey of physics-informed machine learning
Meng, C., Griesemer, S., Cao, D., Seo, S., & Liu, Y. (2025) · Machine learning for computational science and engineering · doi:10.1007/s44379-025-00016-0 · Cited by 329 · see citing works
“Examples of real-world object-centric data include trajectory data from multi-agent sys-123 tems [114–116], position and velocity data from networks of motion sensors [115, 117], and molecule data [54].”
Hyperbolic Deep Neural Networks: A Survey
Peng, W., Varanka, T., Mostafa, A., Shi, H., & Zhao, G. (2021) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2021.3136921 · Cited by 220 · see citing works
“Work [48] shares the same idea to construct spatial temporal graph convolutional networks [144] in the hyperbolic space and applied this for dynamic graph sequences.”
BackgroundGaitgraph: Graph Convolutional Network for Skeleton-Based Gait Recognition
Teepe, T., Khan, A. R., Gilg, J. J., Herzog, F., Hörmann, S., & Rigoll, G. (2021) · International Conference on Information Photonics · doi:10.1109/icip42928.2021.9506717 · Cited by 324
“Inspired by the success of GCNs in skeleton-based action recognition [9, 10], we adapted the methods to the gait recognition task.”
MethodUnsupervised Domain Adaptation of Object Detectors: A Survey
Oza, P., Sindagi, V. A., VS, V., & Patel, V. M. (2024) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2022.3217046 · Cited by 313 · see citing works
Skeleton Aware Multi-modal Sign Language Recognition
Jiang, S., Sun, B., Wang, L., Bai, Y., Li, K., & Fu, Y. (2021) · 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) · doi:10.1109/cvprw53098.2021.00380 · Cited by 309 · see citing works
Yadav, S. K., Tiwari, K., Pandey, H. M., & Akbar, S. (2021) · Knowledge-Based Systems · doi:10.1016/j.knosys.2021.106970 · Cited by 328 · see citing works
Computing Graph Neural Networks: A Survey from Algorithms to Accelerators
Abadal, S., Jain, A., Guirado, R., López-Alonso, J., & Alarcón, E. (2021) · ACM Computing Surveys · doi:10.1145/3477141 · Cited by 293 · see citing works
Graph Stacked Hourglass Networks for 3D Human Pose Estimation
Xu, T., & Takano, W. (2021) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr46437.2021.01584 · Cited by 306 · see citing works
“Moreover, our hourglass module can be seen as a wellintegrated, high-performance graph convolution module, indicating that our hourglass module is general and can be easily extended to other tasks using graph convolution, such as action recognition [46, 20], motion prediction [21], etc.”
MethodHierarchically Decomposed Graph Convolutional Networks for Skeleton-Based Action Recognition
Lee, J., Lee, M., Lee, D., & Lee, S. (2023) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51070.2023.00958 · Cited by 333 · see citing works
“However, existing GCN-based methods [35, 24, 25, 4, 2] have the following limitations.”
BackgroundMethodInfluentialGraph representation learning in biomedicine and healthcare
Li, M. M., Huang, K., & Zitnik, M. (2022) · Nature Biomedical Engineering · doi:10.1038/s41551-022-00942-x · Cited by 294 · see citing works
“Specifically, the inductive ability of graph convolutional networks (GCNs) to generalize to data points unseen during model training, and to generate new data points from scratch by decoding latent representation from the embedding space, has enabled the discovery of new molecules, interactions and functions 100,105,1…”
InfluentialGraph-Based Deep Learning for Medical Diagnosis and Analysis: Past, Present and Future
Ahmedt‐Aristizabal, D., Armin, M. A., Denman, S., Fookes, C., & Petersson, L. (2021) · Sensors · doi:10.3390/s21144758 · Cited by 287 · see citing works
“ST-GCN model: ST-GCN are popular for solving problems that base predictions on graph-structured time series [63].”
BackgroundMethodUAV-Human: A Large Benchmark for Human Behavior Understanding with Unmanned Aerial Vehicles
Li, T., Liu, J., Zhang, W., Ni, Y., Wang, W., & Li, Z. (2021) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr46437.2021.01600 · Cited by 298 · see citing works
“Here we evaluate different state-of-the-art methods [39, 31, 30, 19, 6] on skeleton modality of the UAV-Human.”
MethodMSR-GCN: Multi-Scale Residual Graph Convolution Networks for Human Motion Prediction
Dang, L., Nie, Y.-W., Long, C., Zhang, Q., & Li, G.-Q. (2021) · IEEE International Conference on Computer Vision · doi:10.1109/iccv48922.2021.01127 · Cited by 284 · see citing works
“, biological gene, point cloud, human social network [48], and human motion prediction for the graph-structure nature of the human skeleton.”
BackgroundCross-Domain Detection via Graph-Induced Prototype Alignment
Xu, M., Wang, H., Ni, B., Tian, Q. C., & Zhang, W. (2020) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr42600.2020.01237 · Cited by 274 · see citing works
“, action recognition [50], person Re-ID [51], video understanding [45, 52] and point cloud learning [23].”
BackgroundSpatio-Temporal Graph for Video Captioning With Knowledge Distillation
Pan, B., Cai, H., Huang, D.-A., Lee, K.-H., Gaidon, A., Adeli, E., & Niebles, J. C. (2020) · 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr42600.2020.01088 · Cited by 281 · see citing works
“We follow [47] to add in the residual connection and use ReLU as the activation function.”
BackgroundMethodSpatiotemporal Co-Attention Recurrent Neural Networks for Human-Skeleton Motion Prediction
Shu, X., Zhang, L., Qi, G.-J., Liu, W., & Tang, J. (2022) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2021.3050918 · Cited by 276 · see citing works
“skeleton [21], [33], [34], [35], [36], [37].”
MethodSpatial Temporal Transformer Network for Skeleton-based Action Recognition
Plizzari, C., Cannici, M., & Matteucci, M. (2020) · ICPR Workshops · doi:10.1007/978-3-030-68796-0_50 · Cited by 274 · see citing works
“ST-GCN [16] 85.”
MethodInfluentialPREDICT & CLUSTER: Unsupervised Skeleton Based Action Recognition
Su, K., Liu, X., & Shlizerman, E. (2020) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr42600.2020.00965 · Cited by 244 · see citing works
“In particular, Spatio-Temporal GCN was designed to learn both spatial and temporal patterns from skeleton data [33].”
MethodHuman activity recognition in artificial intelligence framework: a narrative review
Gupta, N., Gupta, S. K., Pathak, R. K., Jain, V., Rashidi, P., & Suri, J. S. (2022) · Artificial Intelligence Review · doi:10.1007/s10462-021-10116-x · Cited by 348 · see citing works
“2020), NTU-RGBD datasets (Yan et al. 2018; Liu et al. 2016), VIRAT 1.”
Spatio-Temporal Graph Routing for Skeleton-Based Action Recognition
Li, B., Li, X., Zhang, Z., & Wu, F. (2019) · AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v33i01.33018561 · Cited by 265 · see citing works
“ST-GCN takes both 3D skeleton and the graph as input and output action class.”
MethodInfluentialDrive&Act: A Multi-Modal Dataset for Fine-Grained Driver Behavior Recognition in Autonomous Vehicles
Martín, M., Roitberg, A., Haurilet, M., Horne, M., Reiß, S., Voit, M., & Stiefelhagen, R. (2019) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2019.00289 · Cited by 199 · see citing works
“Due to the complex structure of skeletons, popular ways for flattening their representation include recurrent neural networks [32, 28] and graph networks [51].”
BackgroundHow to Build a Graph-Based Deep Learning Architecture in Traffic Domain: A Survey
Ye, J., Zhao, J., Ye, K., & Xu, C. (2020) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2020.3043250 · Cited by 285 · see citing works
“An evolving topological structure [66], [114] is incorporated into the model to capture such dynamic spatial change.”
BackgroundMotion Guided 3D Pose Estimation from Videos
Wang, J., Yan, S.-J., Xiong, Y., & Lin, D.-H. (2020) · European Conference on Computer Vision · doi:10.1007/978-3-030-58601-0_45 · Cited by 257 · see citing works
“Therefore, we represent the skeleton sequence as a spatial temporal graph [38] to maintain their topologies, and aggregating information by an U-shaped graph convolution network (UGCN).”
BackgroundMethodInfluentialSelf-citationGraph Neural Ordinary Differential EquationsPreprint
Poli, M., Massaroli, S., Park, J., Yamashita, A., Asama, H., & Park, J. (2019) · arXiv.org · doi:10.48550/arxiv.1911.07532 · Cited by 256 · see citing works
“GNNs have shown remarkable results in various application areas such as node classification (Atwood and Towsley 2016), graph classification (Yan, Xiong, and Lin 2018) and forecasting (Li et al. 2017; Park and Park 2019) as well as generative tasks (Li et al. 2018).”
BackgroundPYSKL: Towards Good Practices for Skeleton Action Recognition
Duan, H., Wang, J., Chen, K., & Lin, D. (2022) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3503161.3548546 · Cited by 216 · see citing works
Self-citationHOPE-Net: A Graph-Based Model for Hand-Object Pose Estimation
Doosti, B., Naha, S., Mirbagheri, M., & Crandall, D. J. (2020) · Computer Vision and Pattern Recognition · doi:10.1109/cvpr42600.2020.00664 · Cited by 241 · see citing works
“[29] used a graph convolution network for learning a spatial-temporal representation of human body joints for skeleton-based action recognition.”
MethodSymbiotic Graph Neural Networks for 3D Skeleton-Based Human Action Recognition and Motion Prediction
Li, M., Chen, S., Chen, X., Zhang, Y., Wang, Y., & Tian, Q. C. (2021) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2021.3053765 · Cited by 243 · see citing works
“Here we compare Sybio-GNN with a few recently proposed methods: ST-GCN [12], HCN [9], and 2s-AGCN [35].”
BackgroundMethodInfluentialA Comprehensive Study of Deep Video Action RecognitionPreprint
Zhu, Y., Li, X., Liu, C., Zolfaghari, M., Xiong, Y., Wu, C., Zhang, Z., Tighe, J., Manmatha, R., & Li, M. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2012.06567 · Cited by 172 · see citing works
“Although there is previous work on using pose for action recognition [150, 246], P-CNN [23] was one of the first deep learning methods that successfully used pose to improve video action recognition.”
MethodSelf-citationZ-GCNETs: Time Zigzags at Graph Convolutional Networks for Time Series ForecastingPreprint
Chen, Y., Segovia-Domínguez, I., & Gel, Y. R. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2105.04100 · Cited by 96 · see citing works
“In addition to convolution, there has been a surge of interest in applying GCNs to time series forecasting tasks (Yu et al., 2018; Yao et al., 2018; Yan et al., 2018; Guo et al., 2019; Weber et al., 2019; Pareja et al., 2020; Segovia Dominguez et al., 2021).”
BackgroundBoulahia, S. Y., Amamra, A., Madi, M. R., & Daikh, S. (2021) · Machine Vision and Applications · doi:10.1007/s00138-021-01249-8 · Cited by 236 · see citing works
“For instance, the use of the ST-GCN [41] extraction module provides on average 96, 000 features for each skeleton-based action sample.”
MethodInfluential3D Human Action Representation Learning via Cross-View Consistency Pursuit
Li, L., Wang, M., Ni, B., Wang, H., Yang, J., & Zhang, W. (2021) · 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr46437.2021.00471 · Cited by 236 · see citing works
“Further, ST-GCN [57] was proposed to better model the graph structure of skeleton data.”
MethodInfluentialRao, H., Xu, S., Hu, X., Cheng, J., & Hu, B. (2021) · Information Sciences · doi:10.1016/j.ins.2021.04.023 · Cited by 235 · see citing works
“Most existing skeleton-based methods [9]–[11] utilize the supervised learning paradigm to learn action representations, in which a significant number of annotations for action frames or videos is indispensable.”
BackgroundMethodInfluentialHiGCIN: Hierarchical Graph-Based Cross Inference Network for Group Activity Recognition
Yan, R., Xie, L.-X., Tang, J.-H., Shu, X.-B., & Tian, Q. (2020) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2020.3034233 · Cited by 234 · see citing works
“Moreover, some STGNNs [41], [42], [43], [44] have been designed to learn unseen patterns from spatiotemporal graph data (e.g., skeleton-based actions) by combing a convolution and a recurrent operation.”
BackgroundA Survey on 3D Skeleton-Based Action Recognition Using Learning Method
Ren, B., Liu, M., Ding, R., & Liu, H. (2024) · Cyborg and Bionic Systems · doi:10.34133/cbsystems.0100 · Cited by 236 · see citing works
“Nowadays, deep learning techniques utilizing recurrent neural networks (RNNs) [51], convolutional neural networks (CNNs) [52], graph convolutional networks (GCNs), and Transformer-based methods have emerged in this field [53,54].”
BackgroundRichly Activated Graph Convolutional Network for Robust Skeleton-Based Action Recognition
Song, Y., Zhang, Z., Shan, C., & Wang, L. (2020) · IEEE transactions on circuits and systems for video technology (Print) · doi:10.1109/tcsvt.2020.3015051 · Cited by 231 · see citing works
“Compared to STGCN [10], which is the first graph-based model for skeleton action recognition, our method outperforms by 5.8% and 5.3%, respectively.”
BackgroundMethodInfluentialSkeleton-aware networks for deep motion retargeting
Aberman, K., Li, P., Lischinski, D., Sorkine-Hornung, O., Cohen-Or, D., & Chen, B. (2020) · ACM Transactions on Graphics · doi:10.1145/3386569.3392462 · Cited by 230 · see citing works
Learning Dynamic Graph Representation of Brain Connectome with Spatio-Temporal Attention
Kim, B.-H., Ye, J.-C., & Kim, J.-J. (2021) · Neural Information Processing Systems · doi:10.48550/arxiv.2105.13495 · Cited by 230
“Experiment on other models including ST-GCN [15], GIN [23], GCN [24], GraphSAGE [31], and ChebGCN [2] demonstrate exceptional performance of our proposed method for HCP-Task (Table 1).”
MethodInfluentialGraph Neural Architecture Search
Gao, Y., Yang, H., Zhang, P., Zhou, C., & Hu, Y. (2020) · International Joint Conference on Artificial Intelligence · doi:10.24963/ijcai.2020/195 · Cited by 226 · see citing works
“The applications of GNNs span over online recommendation [Wu et al. , 2019b], traffic forecasting [Yu et al. , 2018] and action recognition [Yan et al. , 2018].”
BackgroundDeepGCNs: Making GCNs Go as Deep as CNNs
Li, G., Mueller, M. F., Qian, G., Delgadillo, I. C., Abualshour, A., Thabet, A., & Ghanem, B. (2021) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2021.3074057 · Cited by 218 · see citing works
“Graphs are also used to model human joints for action recognition in video [23], [24].”
MethodMS2L: Multi-Task Self-Supervised Learning for Skeleton Based Action Recognition
Lin, L., Song, S., Yang, W., & Liu, J. (2020) · Proceedings of the 28th ACM International Conference on Multimedia · doi:10.1145/3394171.3413548 · Cited by 223 · see citing works
“30 ST-GCN [43] 81.”
BackgroundMethodInfluentialKim, J., Kim, H., Kim, H., Lee, D., & Yoon, S. (2025) · Artificial Intelligence Review · doi:10.1007/s10462-025-11223-9 · Cited by 334 · see citing works
Music Gesture for Visual Sound Separation
Gan, C., Huang, D. Y., Zhao, H., Tenenbaum, J. B., & Torralba, A. B. (2020) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr42600.2020.01049 · Cited by 197 · see citing works
“Similar to [53], we start by constructing a undirected spatial-temporal graph G = {V,E} on a human skeleton sequence.”
MethodSpatio-Temporal Graph Convolution for Resting-State fMRI Analysis
Gadgil, S. U., Zhao, Q., Pfefferbaum, A., Sullivan, E. V., Adeli, E., & Pohl, K. M. (2020) · Lecture notes in computer science · doi:10.1007/978-3-030-59728-3_52 · Cited by 210 · see citing works
“To address these issues, we suggest deep neural networks to incorporate spatio-temporal convolution on functional connectivity graphs, i.e., spatio-temporal Graph Convolution (ST-GC) [24,25].”
BackgroundMethodInfluentialGraph Embedded Pose Clustering for Anomaly Detection
Markovitz, A., Sharir, G., Friedman, I., Zelnik-Manor, L., & Avidan, S. (2019) · Computer Vision and Pattern Recognition · doi:10.1109/cvpr42600.2020.01055 · Cited by 222 · see citing works
“[27] and Yu et al .”
BackgroundMethodInfluentialClass-Incremental Continual Learning Into the eXtended DER-Verse
Boschini, M., Bonicelli, L., Buzzega, P., Porrello, A., & Calderara, S. (2022) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2022.3206549 · Cited by 133 · see citing works
“We read the raw data as done by the official codebase for [76], which represents each sample as a S × T × J × B tensor, where S = 3 is the number of dimensions in space, T = 300 is the number of frames per sequence, J = 25 is the number of tracked skeletal compute the rescaling factor M ← save new…”
CATER: A diagnostic dataset for Compositional Actions and TEmporal ReasoningPreprint
Girdhar, R., & Ramanan, D. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1910.04744 · Cited by 74 · see citing works
Space-Time-Separable Graph Convolutional Network for Pose Forecasting
Sofianos, T., Sampieri, A., Franco, L., & Galasso, F. (2021) · IEEE International Conference on Computer Vision · doi:10.1109/iccv48922.2021.01102 · Cited by 214 · see citing works
“These have mostly been hand-designed, mainly leveraging the natural structure of the kinematic tree [25, 8, 51], and encoded via Graph Convolu-tional Networks (GCN) [27].”
InfluentialAdaptive Context-Aware Multi-Modal Network for Depth Completion
Zhao, S., Gong, M., Fu, H., & Tao, D. (2021) · IEEE Transactions on Image Processing · doi:10.1109/tip.2021.3079821 · Cited by 194 · see citing works
“To model the graph data efficiently, Graph Models have been applied on various computer vision tasks, such as action recognition [64, 65], point cloud analysis [66, 23], few-shot image classification [67, 68], and person reidentification [69].”
MethodUncertainty-Aware Score Distribution Learning for Action Quality Assessment
Tang, Y., Ni, Z., Zhou, J., Zhang, D., Lu, J., Wu, Y., & Zhou, J. (2020) · 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr42600.2020.00986 · Cited by 209 · see citing works
“ST-GCN [16,38] 0.”
Graph Neural Networks: Taxonomy, Advances, and Trends
Zhou, Y. P., Zheng, H., Huang, X., Hao, S., Li, D., & Zhao, J. (2022) · ACM Transactions on Intelligent Systems and Technology · doi:10.1145/3495161 · Cited by 170 · see citing works
“Spatio-temporal graphs can be used to model traffic networks [59, 71] and skeleton networks [55, 57].”
MethodMulti-Label Classification with Label Graph Superimposing
Wang, Y., He, D.-L., Li, F., Long, X., Zhou, Z., Ma, J.-W., & Wen, S.-L. (2019) · AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v34i07.6909 · Cited by 208 · see citing works
“For instance, it was leveraged in (Yan, Xiong, and Lin 2018; Gao et al. 2018) to model the relationship of skeletons of humans bodies for human action recognition and knowledge-aware GCN was applied for zero-shot video classification in (Gao, Zhang, and Xu 2019).”
MethodDiffPose: Toward More Reliable 3D Pose Estimation
Gong, J. L., Foo, L. G., Fan, Z., Ke, Q., Rahmani, H., & Liu, J. (2023) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52729.2023.01253 · Cited by 194 · see citing works
“Unlike these generation tasks [20, 25] that often start the generation process from random noise, our pose estimation process starts from an estimated 2D pose with uncertainty and indeterminacy in 3D space, where the uncertainty distribution differs for each pose and can also be irregular and difficult to characterize.”
BackgroundChen, C., Wu, Y., Dai, Q., Zhou, H.-Y., Xu, M., Yang, S., Han, X., & Yu, Y. (2024) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2024.3445463 · Cited by 177 · see citing works
“Considering the human skeleton has a graph structure with natural joint connections, Yan et al. [99] propose an ST-GCN network that first connects joints in the same frame according to the natu-ral connectivity in the human skeleton, and then connects the same joints in two consecutive frames to…”
BackgroundMethodInfluentialSpatial-Temporal Graph Convolutional Network for Video-Based Person Re-Identification
Yang, J., Zheng, W.-S., Yang, Q., Chen, Y.-C., & Tian, Q. (2020) · Computer Vision and Pattern Recognition · doi:10.1109/cvpr42600.2020.00335 · Cited by 198 · see citing works
“In recent years, graph convolutional networks (GCNs) and its variants [20, 5, 14, 18] have been successfully applied to some tasks in computer vision, such as skeleton-based action recognition [62], video classication [52], and multi-label image recognition [7] due to its strong ability of relations modeling.”
BackgroundCross-Node Federated Graph Neural Network for Spatio-Temporal Data Modeling
Meng, C.-Z., Rambhatla, S., & Liu, Y. (2021) · Knowledge Discovery and Data Mining · doi:10.1145/3447548.3467371 · Cited by 197 · see citing works
“sensors, wearable devices and the Internet of Things (IoT) devices) is critical for various applications including traffic flow prediction [18, 32], forecasting [4, 24], and user activity detection [20, 29].”
BackgroundKumar, P., Chauhan, S., & Awasthi, L. K. (2023) · Archives of Computational Methods in Engineering · doi:10.1007/s11831-023-09986-x · Cited by 220 · see citing works
“Many GNN and GCN-based HAR methods [207, 208] have been proposed because the skeleton data can be represented as a graph with edges and nodes.”
Improving Action Segmentation via Graph-Based Temporal Reasoning
Huang, Y., Sugano, Y., & Sato, Y. (2020) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr42600.2020.01404 · Cited by 160 · see citing works
“Since then, GCNs have shown convincing successes in modeling relations [2, 15, 22] and thus are widely applied in multiple research tasks like semisupervised learning [38], image captioning [63], skeletonbased action recognition [62], and video action recognition [58, 60, 65, 66].”
BackgroundShaik, T., Tao, X., Li, L., Xie, H., & Velasquez, J. D. (2023) · Information Fusion · doi:10.1016/j.inffus.2023.102040 · Cited by 213 · see citing works
“Graph-based algorithms, such as graph convolutional networks (GCNs) or graph regularized non-negative matrix factorization (GNMF), can then be applied to capture the dependencies and interactions between modalities [97].”
MethodSignBERT+: Hand-Model-Aware Self-Supervised Pre-Training for Sign Language Understanding
Hu, H., Zhao, W., Zhou, W., & Li, H. (2023) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2023.3269220 · Cited by 154 · see citing works
“Spatially, following [24], we adopt random moving augmentation to simulate spatial disturbance induced by rotation, translation and scaling factors.”
MethodInfluentialSkeleton-Contrastive 3D Action Representation Learning
Thoker, F. M., Doughty, H., & Snoek, C. G. M. (2021) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3474085.3475307 · Cited by 159 · see citing works
- Authors
- Yan, S., Xiong, Y., & Lin, D.
- Affiliations
- Chinese University of Hong Kong (HK)
- Venue
- Proceedings of the AAAI Conference on Artificial Intelligence
- Published
- 2018-04-27
- Volume
- 32
- Issue
- 1
- ISSN
- 2159-5399 (ISSN-L)
- 2374-3468
- OpenAlex
- W2963076818
- Cited by
- 5,410
- References
- 32
- Citations by year
- 2026: 552 · 2025: 895 · 2024: 915 · 2023: 915 · 2022: 755 · 2021: 742 · 2020: 414 · 2019: 193 · 2018: 29
- FWCI
- 103.84
- Type
- conference paper
- Language
- en
- Open access
- diamond
- doi.org
- Locations
- Proceedings of the AAAI Conference on Artificial Intelligence (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- AAAI Publications (The Association for the Advancement of Artificial Intelligence (AAAI)) (submittedVersion)
- arXiv (Cornell University)
- Funders
- Chinese University of Hong Kong
- SenseTime 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 › Engineering › Control and Systems Engineering › Human Motion and Animation
- Keywords
- skeleton-based action recognition, spatio-temporal graph convolutional network, graph convolutional networks, temporal modeling, spatial modeling, generalization
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
