Pith. sign in

REVIEW 3 major objections 5 minor 58 references

Adaptive Hyper-Graph Convolution Network for Skeleton-based Human Action Recognition with Virtual Connections

T0 review · 3 major / 5 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read The paper claims that an adaptive hyper-graph convolutional network with learnable virtual connections achieves state-of-the-art accuracy on three skeleton-action benchmarks using fewer parameters than comparable GCN and Transformer…

desk verdict The adaptive hypergraph idea is real, but the virtual-joint mechanism is absent from the equations and the SOTA claims are under-powered; this paper needs a serious revision before the headline numbers can be trusted. read the letter →

arxiv 2411.14796 v3 pith:7HKFMREB submitted 2024-11-22 cs.CV cs.LG

Youwei Zhou , Tianyang Xu , Cong Wu , Xiaojun Wu , Josef Kittler This is my paper · ORCID
classification cs.CVcs.LG
keywords skeleton-basedhumanactionrecognitionhyper-graphconvolutionadaptivenon-uniformvirtualconnectionsgraphconvolutionalnetworksNTURGB+DNW-UCLAmulti-head
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 claims that treating the human skeleton as a hyper-graph, where each hyper-edge can link more than two joints at once, captures action-defining multi-joint synergies better than the binary edges used by standard graph convolutional networks. To make that work, it introduces an adaptive, non-uniform hyper-graph construction that is learned during training, so each action gets a topology shaped by its own joint interactions, plus learnable 'virtual hyper-joints' that act as shared semantic carriers connected to all real joints. On the NTU-60, NTU-120, and NW-UCLA benchmarks, the resulting Hyper-GCN is reported to outperform all GCN-based and prior hyper-graph-based methods, and to surpass Transformer-based methods on NTU-120, while using fewer parameters than competing approaches. A sympathetic reader would care because it suggests higher-order, action-specific connectivity is a genuinely useful inductive bias for skeleton recognition, not just an incremental architecture change.

What carries the argument

The engine of the method is the Adaptive Non-uniform Hyper-graph (A-NHG) construction combined with Multi-head Hyper-graph Convolution (M-HGC). For each joint, A-NHG measures Euclidean distances in a learned embedding space and keeps only the $K$ nearest hyper-edges, assigning soft probabilities via softmax; because each hyper-edge's membership is not fixed, the hyper-graph is non-uniform and action-specific. M-HGC splits features into eight channel heads, builds one hyper-graph per head, and fuses the normalized hyper-graph incidence matrix with the physical skeleton adjacency matrix under a learnable mixing weight. Learnable hyper-joints, one per layer, are concatenated to real joints and participate in the same hyper-graph convolution, with a divergence loss that keeps them from collapsing into identical tokens.

What would settle it

Retrain Hyper-GCN ten times with different seeds on NTU-120 X-Sub, reporting mean and standard deviation, and compare against the reported baselines; if the gap to the second-best method is smaller than the standard deviation, the superiority claim does not hold. Alternatively, choose $K$ and the number of hyper-joints on a validation split disjoint from the test split and check whether the reported test accuracy persists.

Watch

Extended reading notes

Core claim

The central claim is that adaptive non-uniform hyper-graphs, together with virtual hyper-joints, let a lightweight graph network model multi-vertex coordination directly, yielding state-of-the-art accuracy on three standard skeleton-action benchmarks. As reported, the base version uses only 1.1M parameters and achieves 93.3/97.4% on NTU-60 X-Sub/X-View, 90.5/91.7% on NTU-120 X-Sub/X-Set, and 97.2% on NW-UCLA; the large version, at 2.3M parameters, reaches 93.7/97.8, 90.9/92.0, and 97.6, taking first place on four of the five benchmarks in the comparison table. The paper argues that these results demonstrate that hyper-graph convolution with adaptive, non-uniform edges is a more efficient way to aggregate skeleton semantics than the binary adjacency matrices of prior GCNs and the heavier attention maps of Transformers.

Load-bearing premise

The state-of-the-art claim rests on single-run accuracy comparisons in which the key hyperparameters ($K$ and the number of hyper-joints) were selected on the same NTU-120 X-Sub split used to report the final result.

Editorial extensions

If this is right

  • If the reported accuracy holds, a 1.1M-parameter base model outperforms all compared GCN and hyper-graph baselines on NTU-120, supporting the claim that multi-vertex aggregation is more efficient than binary-edge message passing.
  • The large version's first-place finish on four of five benchmarks suggests that scaling the hyper-graph architecture, rather than adding heavier attention, is a viable path to better skeleton action recognition.
  • The ablation results imply that non-uniform hyper-graphs (with $K=9$) beat uniform ones, and that adding exactly three virtual hyper-joints helps most; more hyper-joints degrade accuracy.
  • Because the hyper-joints are learnable parameters shared across frames and layers, the method positions them as generalizable semantic carriers, analogous to class tokens in Transformers.

Reading between the lines

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

  • Extending beyond the paper, the same adaptive hyper-edge construction could be applied to other structured-sequence tasks, such as multi-agent motion forecasting or gesture recognition from mocap, where interactions among more than two entities are known to matter.
  • The virtual hyper-joint mechanism suggests a concrete bridge to Transformer class tokens: one could initialize hyper-joints from a pretrained global representation and measure whether they absorb dataset-level action semantics, a test the paper does not run.
  • A testable extension would be cross-dataset transfer: train Hyper-GCN on NTU-120 and finetune on a smaller skeleton dataset to see whether the learned hyper-graph patterns generalize beyond the benchmark's action vocabulary.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. The paper proposes Hyper-GCN, a skeleton-based action recognition architecture that replaces fixed binary graph topology with an adaptive non-uniform hyper-graph. The method constructs K-nearest hyper-edges per joint in a learned embedding space, applies an 8-head multi-head hyper-graph convolution, and introduces learnable "hyper-joints" with a divergence loss as virtual connections. The architecture is evaluated with 4-stream ensembles on NTU-RGB+D 60, NTU-RGB+D 120, and NW-UCLA, reporting state-of-the-art accuracy with fewer parameters than most competing GCN and Transformer methods. Ablation studies examine the hyper-parameter K, the number of hyper-joints, and the divergence loss, with visualizations of hyper-edges and t-SNE projections.

Significance. If the empirical claims hold, the paper makes a useful contribution: adaptive non-uniform hyper-graph construction is a plausible extension of edge-based GCNs, the parameter efficiency (1.1M parameters for the base version) is competitive, and the code is released. The core hyper-graph construction and normalization equations in Sections 3.2-3.4 are coherent, and the ablations in Tables 2 and 3 provide evidence for the value of the proposed components. However, the central virtual-connection mechanism is not formalized in the equations, and the SOTA claim rests on single-run accuracy differences without error bars, with hyperparameters selected on a test split. These issues materially weaken confidence in the paper's main claims as currently written.

major comments (3)
  1. [3.4-3.5, Eqs. (9)-(11)] The virtual-connection mechanism that motivates the paper is absent from the formal definition. Eq. (9) builds the incidence matrix H solely from A-NHG applied to \bar{F}^k_in, which are temporally pooled features of real joints only, and Eq. (11) computes F_out = ⊎_{k=1}^8 (\hat{A}_k + α\hat{H}_k) F^k_in P_k with no F_h term anywhere. Section 3.5 states that hyper-joints "are to participate in the hyper-graph convolution" and are "manually connect[ed] to all the physical joints," but no equation implements these statements. Consequently, the improvement in Table 3 when hyper-joints are added cannot be attributed to the proposed virtual-connection mechanism as written. Please extend Eqs. (8)-(11) to show explicitly how F_h enters the incidence matrix and/or the message-passing aggregation, or state precisely how the hyper-joints are used (e.g., only in the auxiliary loss) and adjust the claims accordingly.
  2. [4.3, 4.4, Tables 1-3] The SOTA claim rests on single-run accuracy differences that are small and unaccompanied by error bars or significance tests. For example, Ours (B) exceeds BlockGCN by 0.2 percentage points on NTU120 X-Sub and by 0.3 points on NW-UCLA, and the benefit of adding 3 hyper-joints with M-HGC in Table 3 is 0.2 points (86.9 vs 86.7). In addition, the hyperparameters K and V_h are selected by ablations on the NTU120 X-Sub split (Tables 2 and 3), and the final results in Table 1 are reported on that same split, which can inflate the reported accuracy through test-set tuning. Please provide multiple runs with mean and standard deviation, and either select hyperparameters on a validation split or demonstrate that the conclusions are stable across seeds.
  3. [4.3, Table 1] The claim that the base version "comprehensively outperforms all GCN-based and HGCN-based SOTA and surpasses the Transformer-based SOTA on the NTU120" is internally consistent, but the paper does not report a per-benchmark comparison against SkateFormer, which is higher than Ours (B) on NTU60 X-Sub, NTU60 X-View, and NW-UCLA in Table 1. Since the abstract and conclusion state SOTA performance on three public datasets, please qualify the claim to specify on which benchmarks and configurations the superiority holds, and state whether the large version or the 4-stream ensemble is being referenced.
minor comments (5)
  1. [4.4] Section 4.4 begins with "redAll the ablation study..." — the stray "red" artifact should be removed.
  2. [3.4] "LeakeyReLU" should be "LeakyReLU".
  3. [3.3] In the paragraph after Eq. (8), "setseti" should read "the set set_i".
  4. [4.1] "extention" should be "extension".
  5. [Table 1 caption] The caption lists "J, B, JB, and JM" but the modalities column uses "J+B+JM+BM"; the abbreviations should be clarified (likely J, B, JM, BM).

Circularity Check

1 steps flagged · score 3.0 of 10

One benchmark (NTU120 X-Sub) is used both to select hyperparameters K and the number of hyper-joints and to report final SOTA; other benchmarks remain independent, and no definitional circularity is present.

  1. fitted input called prediction [Section 4.4 (Tables 2-3); Section 4.3 (Table 1, Ours (B) NTU120 X-Sub row)]
    "Therefore, setting K = 5 in uniform and K = 9 in Non-uniform achieves the best performance. ... The case of introducing only 3 hyper joints achieves the best performance."

    The hyperparameters K and the number of hyper-joints are selected by maximizing accuracy on the NTU120 X-Sub split in the ablations, and the same split is then reported as the final SOTA result in Table 1. The reported X-Sub accuracy is therefore the maximum of the explored grid over the selection benchmark rather than an independent evaluation of a fixed architecture. This makes the NTU120 X-Sub 'prediction' statistically forced by the selection procedure. This is a mild and localized circularity: NTU60, NTU120 X-Set, and NW-UCLA results are not used for this selection, so the central architecture claim retains independent empirical content.

full rationale

No definitional circularity was found in the formal chain: Eq. (9) constructs H from F_in and Eq. (11) aggregates F_in, so no reported quantity is equal to an input by construction. The hyper-graph normalization and convolution follow standard external definitions [12]. The paper's one self-citation (Graph2Net [50]) is a related-work remark and is not load-bearing. The virtual hyper-joints are asserted to participate in convolution and be manually connected to all real joints, but no equation contains F_h; this is a formal support gap (the claimed mechanism is not visible in Eqs. 9-11), not a circular reduction, so under the hard rules it is not scored as circularity. The only circular element is the test-set hyperparameter selection on NTU120 X-Sub followed by the final comparison on that same split. That is a real but partial form of fitted-input-as-prediction, and it limits the strength of one reported SOTA number without invalidating the other benchmarks.

Assumptions & free parameters 4 free parameters · 5 assumptions · 2 invented entities

The method rests on standard hyper-graph definitions, the ST-GCN skeleton partition, and two design assumptions specific to this paper: that Euclidean proximity in a learned space identifies action-relevant joint groups, and that learnable global joints help recognition. The main free choices are K and the number of hyper-joints, both selected on one of the reported test splits.

free parameters (4)
  • K (number of nearest hyper-edges per joint) = 9
    Selected via ablation on NTU120 X-Sub (Table 2); the same split is used in the final comparison, so it is tuned on the evaluation set.
  • Vh (number of hyper-joints per layer) = 3
    Selected via ablation on NTU120 X-Sub (Table 3); again the selection set matches the reported test split.
  • Number of heads in M-HGC = 8
    Fixed design choice; no ablation shown for head count, and each head has its own hyper-graph, increasing capacity.
  • Ensemble stream configuration = J, B, JM, BM (4 streams)
    Final results use the standard four-stream fusion; the joint-only numbers are reported only for ablations, not in Table 1, so the headline accuracy includes modality engineering beyond the proposed module.
assumptions (5)
  • standard math Hyper-graph incidence and normalization (Eqs. 3-6) correctly generalize graph convolution to multi-vertex edges.
    Taken from prior hyper-graph literature [2, 12]; the paper uses D_v^-1 H W D_e^-1 H^T as the normalized incidence operator.
  • domain assumption The three-subset physical topology partition (identity, centrifugal, centripetal) is a valid encoding of skeleton bone relations.
    Adopted from ST-GCN [52]; the paper fuses this physical adjacency with learned hyper-graphs without questioning the partition.
  • ad hoc to paper Euclidean distance in a learned embedding subspace reveals action-relevant multi-joint relations for K-nearest hyper-edge assignment.
    This is the core premise of A-NHG (Eqs. 7-8); it is validated only through downstream accuracy, with no independent evidence that Euclidean proximity in the subspace corresponds to joint synergy.
  • ad hoc to paper Learnable hyper-joints shared across frames and layers can store global semantic information and improve recognition.
    Motivated by an analogy to class tokens and shadow puppetry; the divergence loss is introduced to prevent collapse, but the representational benefit is assumed.
  • domain assumption Single-run benchmark accuracy is stable enough for the reported method comparisons.
    No variance, confidence intervals, or multiple seeds are reported; all conclusions assume Table 1 differences are not noise.
invented entities (2)
  • Hyper-joints (virtual joints)
    purpose: Learnable vectors in each layer, shared over frames and connected to all physical joints, acting as global semantic carriers to offload long-range information from real joints.
    The only evidence is internal ablation (Table 3) and cosine-matrix visualizations; there is no external falsifiable prediction, so the entity has no independent confirmation.
  • Adaptive non-uniform hyper-edges
    purpose: Per-sample multi-joint connections formed by K-nearest selection in a learned feature space, intended to capture collaborative joint patterns beyond pairwise bones.
    The hyper-edges are a dynamic construction inside the network; their benefit is shown only by internal accuracy gain and visualizations, with no separate testable handle.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Adaptive Hyper-Graph Convolution Network for Skeleton-based Human Action Recognition with Virtual Connections." pith.science (2026). https://pith.science/paper/7HKFMREB

@misc{pith2026241114796,
  author       = {Pith},
  title        = {Pith review of: Adaptive Hyper-Graph Convolution Network for Skeleton-based Human Action Recognition with Virtual Connections},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/7HKFMREB}},
  note         = {Machine review of arXiv:2411.14796}
}
read the original abstract

The shared topology of human skeletons motivated the recent investigation of graph convolutional network (GCN) solutions for action recognition. However, most of the existing GCNs rely on the binary connection of two neighboring vertices (joints) formed by an edge (bone), overlooking the potential of constructing multi-vertex convolution structures. Although some studies have attempted to utilize hyper-graphs to represent the topology, they rely on a fixed construction strategy, which limits their adaptivity in uncovering the intricate latent relationships within the action. In this paper, we address this oversight and explore the merits of an adaptive hyper-graph convolutional network (Hyper-GCN) to achieve the aggregation of rich semantic information conveyed by skeleton vertices. In particular, our Hyper-GCN adaptively optimises the hyper-graphs during training, revealing the action-driven multi-vertex relations. Besides, virtual connections are often designed to support efficient feature aggregation, implicitly extending the spectrum of dependencies within the skeleton. By injecting virtual connections into hyper-graphs, the semantic clues of diverse action categories can be highlighted. The results of experiments conducted on the NTU-60, NTU-120, and NW-UCLA datasets demonstrate the merits of our Hyper-GCN, compared to the state-of-the-art methods. The code is available at https://github.com/6UOOON9/Hyper-GCN.

Figures

Figures reproduced from arXiv: 2411.14796 by the authors.

Figure 1
Figure 1. Illustration of our Hyper-GCN. In (a), dotted lines represent the virtual connections. Each coloured part containing multiple joints represents a hyper-graph with hyper-edges. (b) pro￾vides an intuitive comparison between normal graph convolution and hyper-graph convolution operations with the same connectiv￾ity degree. a series of 2D or 3D coordinates, which can easily be col￾lected by low-cost depth sensors or obt… view at source ↗
Figure 2
Figure 2. The framework of our Hyper-GCN. Part (a) represents the Multi-head Hyper-graph Convolution (M-HGC) module. Part (b) represents the process of constructing an adaptive hyper-graph. Part (c) represents the architecture of Hyper-GCN. Fp represents the features of real joints. Fh represents the hyper-joints which are learnable parameters. Aˆ represents the physical topology. represents the number of hyper-edges. Given v… view at source ↗
Figure 3
Figure 3. Instance of the incidence matrix construction by A￾NHG with hyper-parameter K = 3. N represents the number of joints. (a) represents the hyper-edge sets are null. (b) represents that measure the distance between 1-st joint and other joints. (c) represents that retrain K-nearest hyper-edges which contain 1-st joint. (d) represents the final state. where Hˆ ∈ R N×N represents the normalised incidence ma￾trix for hyper… view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Visualisation of hyper-graph in selected actions. For ease of observation, the joints contained in the selected hyper-edge are highlighted in red. The results are reported in [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]
Figure 6
Figure 6. Figure 6: t-SNE visualisation of Baseline and Hyper-GCN. Dif￾ferent colours represent different joints. Part (a) and (b) reprsent the output features in Baseline and Hyper-GCN. the best performance. As our hyper joints are learned from a large amount of data, involving a large n…
Figure 5
Figure 5. Figure 5: Comparison of the cosine matrices of hyper-joints. Part (a) and (b) represents w/o Divergence Loss and w/ Divergence Loss with 3 hyper-joints. It is calculated by Eqn. (12). action represented by the hyper-edge. As the K decreases, it is difficult for the hyper-edge to…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

58 extracted references · 54 canonical work pages

  1. [1]

    Skeleton-based action recognition using sparse spatio- temporal gcn with edge effective resistance

    Tasweer Ahmad, Lianwen Jin, Luojun Lin, and GuoZhi Tang. Skeleton-based action recognition using sparse spatio- temporal gcn with edge effective resistance. Neurocomput- ing, 423:389–398, 2021. 3

    work page 2021

  2. [2]

    Hypergraph convolution and hypergraph attention

    Song Bai, Feihu Zhang, and Philip HS Torr. Hypergraph convolution and hypergraph attention. Pattern Recognition, 110:107637, 2021. 2

    2021

  3. [3]

    Channel-wise topology refinement graph convolution for skeleton-based action recognition

    Yuxin Chen, Ziqi Zhang, Chunfeng Yuan, Bing Li, Ying Deng, and Weiming Hu. Channel-wise topology refinement graph convolution for skeleton-based action recognition. In ICCV, pages 13339–13348, 2021. 1, 2, 3, 5, 6, 7

    work page 2021

  4. [4]

    Decoupling gcn with dropgraph module for skeleton-based action recognition

    Ke Cheng, Yifan Zhang, Congqi Cao, Lei Shi, Jian Cheng, and Hanqing Lu. Decoupling gcn with dropgraph module for skeleton-based action recognition. In ECCV, page 536–553, Berlin, Heidelberg, 2020. Springer-Verlag. 3, 7

    work page 2020

  5. [5]

    Skeleton-based action recogni- tion with shift graph convolutional network

    Ke Cheng, Yifan Zhang, Xiangyu He, Weihan Chen, Jian Cheng, and Hanqing Lu. Skeleton-based action recogni- tion with shift graph convolutional network. InCVPR, pages 180–189, 2020. 3

    work page 2020

  6. [6]

    In- fogcn: Representation learning for human skeleton-based ac- tion recognition

    Hyung-Gun Chi, Myoung Hoon Ha, Seunggeun Chi, Sang Wan Lee, Qixing Huang, and Karthik Ramani. In- fogcn: Representation learning for human skeleton-based ac- tion recognition. In CVPR, pages 20154–20164, 2022. 1, 2, 3, 6, 7

    work page 2022

  7. [7]

    P-cnn: Pose-based cnn features for action recognition

    Guilhem Ch ´eron, Ivan Laptev, and Cordelia Schmid. P-cnn: Pose-based cnn features for action recognition. In ICCV, pages 3218–3226, 2015. 2

    work page 2015

  8. [8]

    Skateformer: skeletal- temporal transformer for human action recognition

    Jeonghyeok Do and Munchurl Kim. Skateformer: skeletal- temporal transformer for human action recognition. In ECCV, pages 401–420. Springer, 2024. 3, 7

    work page 2024

Show all 58 references
  1. [9]

    An image is worth 16x16 words: Trans- formers for image recognition at scale

    Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Syl- vain Gelly, et al. An image is worth 16x16 words: Trans- formers for image recognition at scale. arXiv preprint a...

    2010 arXiv

  2. [10]

    Hierarchical recur- rent neural network for skeleton based action recognition

    Yong Du, Wei Wang, and Liang Wang. Hierarchical recur- rent neural network for skeleton based action recognition. In CVPR, pages 1110–1118, 2015. 1

    2015

  3. [11]

    Multi-scale spatial temporal graph neural network for skeleton-based action recognition

    Dong Feng, Zhongcheng Wu, Jun Zhang, and Tingting Ren. Multi-scale spatial temporal graph neural network for skeleton-based action recognition. IEEE Access, 9:58256– 58265, 2021. 3, 7

    2021

  4. [12]

    Hgnn+: General hypergraph neural networks

    Yue Gao, Yifan Feng, Shuyi Ji, and Rongrong Ji. Hgnn+: General hypergraph neural networks. TPAMI, 45(3):3181– 3199, 2023. 2, 3

    2023

  5. [13]

    Hypergraph neural network for skeleton-based action recog- nition

    Xiaoke Hao, Jie Li, Yingchun Guo, Tao Jiang, and Ming Yu. Hypergraph neural network for skeleton-based action recog- nition. TIP, 30:2263–2275, 2021. 2, 3, 7

    2021

  6. [14]

    Multi-channel network: Constructing efficient gcn baselines for skeleton-based action recognition

    Ruijie Hou, Zhihao Wang, Ruimin Ren, Yang Cao, and Zhao Wang. Multi-channel network: Constructing efficient gcn baselines for skeleton-based action recognition. Computers & Graphics, 110:111–117, 2023. 3

    2023

  7. [15]

    Part-level graph convolutional network for skeleton- based action recognition

    Linjiang Huang, Yan Huang, Wanli Ouyang, and Liang Wang. Part-level graph convolutional network for skeleton- based action recognition. In AAAI, pages 11045–11052,

  8. [16]

    Gesture interaction in virtual reality: A low-cost machine learning system and a qualitative assessment of effectiveness of se- lected gestures vs

    Cloe Huesser, Simon Schubiger, and Arzu C ¸¨oltekin. Gesture interaction in virtual reality: A low-cost machine learning system and a qualitative assessment of effectiveness of se- lected gestures vs. gaze and controller interaction. In Inter- national Conference on Human-Comp...

    2021

  9. [17]

    3d convolu- tional neural networks for human action recognition.TPAMI, 35(1):221–231, 2012

    Shuiwang Ji, Wei Xu, Ming Yang, and Kai Yu. 3d convolu- tional neural networks for human action recognition.TPAMI, 35(1):221–231, 2012. 2

    2012

  10. [18]

    Dynamic hypergraph neural networks

    Jianwen Jiang, Yuxuan Wei, Yifan Feng, Jingxuan Cao, and Yue Gao. Dynamic hypergraph neural networks. In IJCAI, pages 2635–2641, 2019. 2

    2019

  11. [19]

    Multilevel hypergraph partitioning: Application in vlsi domain

    George Karypis, Rajat Aggarwal, Vipin Kumar, and Shashi Shekhar. Multilevel hypergraph partitioning: Application in vlsi domain. In Proceedings of the 34th annual Design Au- tomation Conference, pages 526–529, 1997. 2

    1997

  12. [20]

    Ddgcn: A dynamic directed graph convolutional network for action recognition

    Matthew Korban and Xin Li. Ddgcn: A dynamic directed graph convolutional network for action recognition. In ECCV, page 761–776, Berlin, Heidelberg, 2020. Springer- Verlag. 3

    2020

  13. [21]

    Semantic human activity annotation tool using skeletonized surveillance videos

    Bokyung Lee, Michael Lee, Pan Zhang, Alexander Tessier, and Azam Khan. Semantic human activity annotation tool using skeletonized surveillance videos. In Proceedings of the 2019 ACM International Symposium on Wearable Com- puters, pages 312–315, 2019. 1

    2019

  14. [22]

    Leveraging spatio- temporal dependency for skeleton-based action recognition

    Jungho Lee, Minhyeok Lee, Suhwan Cho, Sungmin Woo, Sungjun Jang, and Sangyoun Lee. Leveraging spatio- temporal dependency for skeleton-based action recognition. In ICCV, pages 10221–10230, 2023. 3

    2023

  15. [23]

    Hierarchically decomposed graph convolutional net- works for skeleton-based action recognition

    Jungho Lee, Minhyeok Lee, Dogyoon Lee, and Sangyoun Lee. Hierarchically decomposed graph convolutional net- works for skeleton-based action recognition. In ICCV, pages 10410–10419, 2023. 2, 3, 5, 7

    2023

  16. [24]

    Actional-structural graph convolutional networks for skeleton-based action recognition

    Maosen Li, Siheng Chen, Xu Chen, Ya Zhang, Yanfeng Wang, and Qi Tian. Actional-structural graph convolutional networks for skeleton-based action recognition. In CVPR, pages 3595–3603, 2019. 3

    2019

  17. [25]

    Kot, and Gang Wang

    Jun Liu, Amir Shahroudy, Dong Xu, Alex C. Kot, and Gang Wang. Skeleton-based action recognition using spatio- temporal lstm network with trust gates. TPAMI, 40(12): 3007–3021, 2018. 1

    2018

  18. [26]

    Jun Liu, Amir Shahroudy, Mauricio Perez, Gang Wang, Ling-Yu Duan, and Alex C. Kot. Ntu rgb+d 120: A large-scale benchmark for 3d human activity understanding. TPAMI, 42(10):2684–2701, 2020. 1, 2, 6

    2020

  19. [27]

    Disentangling and unifying graph con- volutions for skeleton-based action recognition

    Ziyu Liu, Hongwen Zhang, Zhenghao Chen, Zhiyong Wang, and Wanli Ouyang. Disentangling and unifying graph con- volutions for skeleton-based action recognition. In CVPR, pages 140–149, 2020. 3, 7

    2020

  20. [28]

    Nikam and Aarti G

    Ashish S. Nikam and Aarti G. Ambekar. Sign language recognition using image based hand gesture recognition techniques. In 2016 Online International Conference on Green Engineering and Technologies, pages 1–5, 2016. 1

    2016

  21. [29]

    Iot-based vibration sensor data collection and emergency detection classification using long short term memory

    Cosmas Ifeanyi Nwakanma, Fabliha Bushra Islam, Mareska Pratiwi Maharani, Dong-Seong Kim, and Jae- Min Lee. Iot-based vibration sensor data collection and emergency detection classification using long short term memory. In International Conference on Artificial Intelli- gence i...

    2021

  22. [30]

    A skeleton-based approach for campus violence detection

    Batyrkhan Omarov, Sergazy Narynov, Zhandos Zhumanov, Aidana Gumar, and Mariyam Khassanova. A skeleton-based approach for campus violence detection. Computers, Mate- rials & Continua, 72(1), 2022. 1

    2022

  23. [31]

    Igformer: Interaction graph transformer for skeleton-based human interaction recognition

    Yunsheng Pang, Qiuhong Ke, Hossein Rahmani, James Bai- ley, and Jun Liu. Igformer: Interaction graph transformer for skeleton-based human interaction recognition. In European Conference on Computer Vision , pages 605–622. Springer,

  24. [32]

    Spa- tial temporal transformer network for skeleton-based ac- tion recognition

    Chiara Plizzari, Marco Cannici, and Matteo Matteucci. Spa- tial temporal transformer network for skeleton-based ac- tion recognition. In Pattern recognition. ICPR international workshops and challenges: virtual event, January 10–15, 2021, Proceedings, Part III, pages 694–701. ...

    2021

  25. [33]

    Spatio- temporal tuples transformer for skeleton-based action recog- nition

    Helei Qiu, Biao Hou, Bo Ren, and Xiaohua Zhang. Spatio- temporal tuples transformer for skeleton-based action recog- nition. arXiv preprint arXiv:2201.02849, 2022. 3

    2022 arXiv

  26. [34]

    Review on synergizing the metaverse and ai-driven synthetic data: enhancing virtual realms and activity recognition in computer vision

    Megani Rajendran, Chek Tien Tan, Indriyati Atmosukarto, Aik Beng Ng, and Simon See. Review on synergizing the metaverse and ai-driven synthetic data: enhancing virtual realms and activity recognition in computer vision. Visual Intelligence, 2(1):27, 2024. 1

    2024

  27. [35]

    Ntu rgb+d: A large scale dataset for 3d human activity anal- ysis

    Amir Shahroudy, Jun Liu, Tian-Tsong Ng, and Gang Wang. Ntu rgb+d: A large scale dataset for 3d human activity anal- ysis. In CVPR, pages 1010–1019, 2016. 1, 2, 6

    2016

  28. [36]

    Two- stream adaptive graph convolutional networks for skeleton- based action recognition

    Lei Shi, Yifan Zhang, Jian Cheng, and Hanqing Lu. Two- stream adaptive graph convolutional networks for skeleton- based action recognition. In CVPR, pages 12018–12027,

  29. [37]

    Skeleton-based action recognition with directed graph neu- ral networks

    Lei Shi, Yifan Zhang, Jian Cheng, and Hanqing Lu. Skeleton-based action recognition with directed graph neu- ral networks. In CVPR, pages 7904–7913, 2019. 3

    2019

  30. [38]

    Decou- pled spatial-temporal attention network for skeleton-based action-gesture recognition

    Lei Shi, Yifan Zhang, Jian Cheng, and Hanqing Lu. Decou- pled spatial-temporal attention network for skeleton-based action-gesture recognition. In Proceedings of the Asian con- ference on computer vision, 2020. 2, 3, 7

    2020

  31. [39]

    Skeleton-based action recognition with multi-stream adap- tive graph convolutional networks

    Lei Shi, Yifan Zhang, Jian Cheng, and Hanqing Lu. Skeleton-based action recognition with multi-stream adap- tive graph convolutional networks. TIP, 29:9532–9545,

  32. [40]

    Two-stream con- volutional networks for action recognition in videos

    Karen Simonyan and Andrew Zisserman. Two-stream con- volutional networks for action recognition in videos. In NIPS, page 568–576, Cambridge, MA, USA, 2014. MIT Press. 2

    2014

  33. [41]

    Constructing stronger and faster baselines for skeleton-based action recognition

    Yi-Fan Song, Zhang Zhang, Caifeng Shan, and Liang Wang. Constructing stronger and faster baselines for skeleton-based action recognition. TPAMI, 45(2):1474–1488, 2022. 3, 7

    2022

  34. [42]

    Skeleton-based human activity recognition for video surveillance

    Ahmed Taha, Hala H Zayed, ME Khalifa, and El-Sayed M El-Horbaty. Skeleton-based human activity recognition for video surveillance. International Journal of Scientific & En- gineering Research, 6(1):993–1004, 2015. 1

    2015

  35. [43]

    Predicting human poses via recurrent attention network

    Jianwei Tang, Jieming Wang, and Jian-Fang Hu. Predicting human poses via recurrent attention network. Visual Intelli- gence, 1(1):18, 2023. 1

    2023

  36. [44]

    Attention is all you need

    Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszko- reit, Llion Jones, Aidan N Gomez, Łukasz Kaiser, and Illia Polosukhin. Attention is all you need. Advances in neural information processing systems, 30, 2017. 2

    2017

  37. [45]

    Differential recurrent neural networks for action recognition

    Vivek Veeriah, Naifan Zhuang, and Guo-Jun Qi. Differential recurrent neural networks for action recognition. In ICCV, pages 4041–4049, 2015. 1

    2015

  38. [46]

    Cross-view action modeling, learning, and recog- nition

    Jiang Wang, Xiaohan Nie, Yin Xia, Ying Wu, and Song- Chun Zhu. Cross-view action modeling, learning, and recog- nition. In CVPR, page 2649–2656, USA, 2014. IEEE Com- puter Society. 2, 6

    2014

  39. [47]

    Iip-transformer: Intra-inter-part transformer for skeleton-based action recognition

    Qingtian Wang, Shuze Shi, Jiabin He, Jianlin Peng, Tingxi Liu, and Renliang Weng. Iip-transformer: Intra-inter-part transformer for skeleton-based action recognition. In Big- Data, pages 936–945, 2023. 2, 3, 7

    2023

  40. [48]

    Dynamic spatial-temporal hypergraph con- volutional network for skeleton-based action recognition

    Shengqin Wang, Yongji Zhang, Hong Qi, Minghao Zhao, and Yu Jiang. Dynamic spatial-temporal hypergraph con- volutional network for skeleton-based action recognition. In 2023 IEEE International Conference on Multimedia and Expo (ICME), pages 2147–2152. IEEE, 2023. 7

    2023

  41. [49]

    Visual- semantic network: a visual and semantic enhanced model for gesture recognition

    Yizhe Wang, Congqi Cao, and Yanning Zhang. Visual- semantic network: a visual and semantic enhanced model for gesture recognition. Visual Intelligence, 1(1):25, 2023. 1

    2023

  42. [50]

    Graph2net: Perceptually-enriched graph learning for skeleton-based ac- tion recognition

    Cong Wu, Xiao-Jun Wu, and Josef Kittler. Graph2net: Perceptually-enriched graph learning for skeleton-based ac- tion recognition. TCSVT, 32(4):2120–2132, 2022. 3

    2022

  43. [51]

    Dynamic semantic-based spa- tial graph convolution network for skeleton-based human ac- tion recognition

    Jianyang Xie, Yanda Meng, Yitian Zhao, Anh Nguyen, Xi- aoyun Yang, and Yalin Zheng. Dynamic semantic-based spa- tial graph convolution network for skeleton-based human ac- tion recognition. AAAI, 38(6):6225–6233, 2024. 1, 3, 7

    2024

  44. [52]

    Spatial tempo- ral graph convolutional networks for skeleton-based action recognition

    Sijie Yan, Yuanjun Xiong, and Dahua Lin. Spatial tempo- ral graph convolutional networks for skeleton-based action recognition. In AAAI, 2018. 2, 3, 5, 7

    2018

  45. [53]

    Two-stream transformer encoders for skeleton-based action recognition

    Hengqi Zhang, Hua Geng, and Geng Yang. Two-stream transformer encoders for skeleton-based action recognition. In International Conference on Computing, Control and In- dustrial Engineering, pages 272–281. Springer, 2021. 2, 3

    2021

  46. [54]

    Semantics-guided neural networks for efficient skeleton-based human action recogni- tion

    Pengfei Zhang, Cuiling Lan, Wenjun Zeng, Junliang Xing, Jianru Xue, and Nanning Zheng. Semantics-guided neural networks for efficient skeleton-based human action recogni- tion. In CVPR, pages 1109–1118, 2020. 3

    2020

  47. [55]

    Stst: Spatial-temporal specialized transformer for skeleton- based action recognition

    Yuhan Zhang, Bo Wu, Wen Li, Lixin Duan, and Chuang Gan. Stst: Spatial-temporal specialized transformer for skeleton- based action recognition. In Proceedings of the 29th ACM International Conference on Multimedia, pages 3229–3237,

  48. [56]

    Learn- ing discriminative representations for skeleton based action recognition

    Huanyu Zhou, Qingjie Liu, and Yunhong Wang. Learn- ing discriminative representations for skeleton based action recognition. In CVPR, pages 10608–10617, 2023. 7

    2023

  49. [57]

    Blockgcn: Redefine topology aware- ness for skeleton-based action recognition

    Yuxuan Zhou, Xudong Yan, Zhi-Qi Cheng, Yan Yan, Qi Dai, and Xian-Sheng Hua. Blockgcn: Redefine topology aware- ness for skeleton-based action recognition. In CVPR, pages 2049–2058, 2024. 2, 3, 7

    2024

  50. [58]

    Selective hypergraph convolutional networks for skeleton-based action recognition

    Yiran Zhu, Guangji Huang, Xing Xu, Yanli Ji, and Fumin Shen. Selective hypergraph convolutional networks for skeleton-based action recognition. In ICMR, page 518–526, New York, NY , USA, 2022. Association for Computing Ma- chinery. 2, 3, 7

    2022

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.