Treffer: Knowledge graph enhanced cross modal generative adversarial network for martial arts motion reconstruction and heritage preservation.
Title:
Knowledge graph enhanced cross modal generative adversarial network for martial arts motion reconstruction and heritage preservation.
Authors:
Source:
Scientific reports [Sci Rep] 2026 Jan 21. Date of Electronic Publication: 2026 Jan 21.
Publication Model:
Ahead of Print
Publication Type:
Journal Article
Language:
English
Journal Info:
Publisher: Nature Publishing Group Country of Publication: England NLM ID: 101563288 Publication Model: Print-Electronic Cited Medium: Internet ISSN: 2045-2322 (Electronic) Linking ISSN: 20452322 NLM ISO Abbreviation: Sci Rep Subsets: MEDLINE
Imprint Name(s):
Original Publication: London : Nature Publishing Group, copyright 2011-
References:
Li, J., Wu, Q. & Zhang, M. Preserving intangible cultural heritage in the digital age: a critical review of AI-based approaches for traditional martial arts. Digit. Herit. 18 (3), 342–361 (2021).
Chen, X., Wang, Y. & Liu, T. Chinese martial arts (Wushu): historical development and contemporary cultural significance. J. Cult. Stud. 45 (2), 187–203 (2020).
Zhang, H., Li, W. & Chen, Y. Challenges and opportunities in the inheritance of traditional martial arts techniques: a systematic analysis. Int. J. Cult. Herit. 15 (4), 478–495 (2022).
Johnson, R. & Smith, K. Traditional knowledge transmission mechanisms in martial arts: limitations and digital transformation paths. J. Cult. Preserv. 12 (3), 267–285 (2021).
Wang, L., Chen, J. & Zhang, Q. Biomechanical analysis of traditional Chinese martial arts: energy flow patterns and technical execution principles. Sports Biomech. Int. 28 (4), 412–429 (2023).
Liu, Y., Zhao, J. & Wu, T. Motion capture technologies for martial arts documentation: a comparative analysis of optical, inertial, and vision-based approaches. IEEE Trans. Vis. Comput. Graph. 29 (2), 1237–1252 (2022).
Chen, H., Wang, Z. & Li, R. Interpretable movement analysis: bridging the gap between motion data and semantic understanding in martial arts. Artif. Intell. Rev. 40 (3), 567–588 (2021).
Wang, Q., Li, S. & Zhang, R. Knowledge graphs for complex domain representation: advances and applications. J. Knowl. Eng. 37 (2), 145–162 (2023).
Yang, M., Chen, X. & Zhao, T. Generative adversarial networks for cultural data synthesis: applications and ethical considerations. IEEE Trans. Neural Netw. Learn. Syst. 33 (9), 4128–4141 (2022).
Zhang, J., Li, H. & Wu, Y. Cross-modal learning frameworks for integrated Understanding of human movement: from perception to generation. Comput. Vis. Image Underst. 227, 103568 (2023).
Brown, M., Johnson, K. & Davis, L. Evolution of martial arts motion documentation: from manual annotation to AI-driven analysis. Int. J. Comput. Vis. 128 (6), 1534–1550 (2020).
Smith, A., Williams, J. & Jones, R. Marker-based motion capture systems for martial arts performance analysis: capabilities and constraints. J. Sports Sci. 39 (5), 518–532 (2021).
Chen, W., Liu, Y. & Zhang, H. Limitations of marker-based systems in capturing authentic martial arts movements: a quantitative assessment. IEEE Sens. J. 22 (7), 6921–6937 (2022).
Park, S., Kim, J. & Lee, M. Markerless vision-based human pose estimation for martial arts documentation: current status and challenges. Comput. Vis. Pattern Recognit. Ann. Rev. 15, 287–304 (2021).
Xu, H., Li, W. & Zhang, T. High-velocity movement tracking in markerless motion capture: performance analysis and optimization strategies. IEEE Trans. Pattern Anal. Mach. Intell. 44 (8), 4562–4578 (2022).
Garcia, J., Martinez, A. & Rodriguez, P. Addressing occlusion challenges in complex martial arts movements: multi-view fusion approaches. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition 3258–3267 (2023).
Wang, L., Chen, R. & Yang, Z. IMU-based motion capture for martial arts: system design and performance evaluation. Sens. Actuators A Phys. 321, 112587 (2021).
Zhang, K., Wu, Y. & Liu, J. Drift compensation in inertial measurement systems for long-duration martial arts performance capture. IEEE Trans. Instrum. Meas. 71, 1–12 (2022).
Li, R., Chen, T. & Wang, M. End-to-end deep learning frameworks for martial arts motion reconstruction: a comparative study. Neural Netw. 158, 174–188 (2023).
Zhao, J., Tao, D. & Wen, L. Transformer-based temporal modeling for complex human movement sequences in traditional martial arts. In Proceedings of the International Conference on Machine Learning 11824–11833 (2022).
Liu, Y., Wang, Z. & Chen, H. Challenges in AI-driven martial arts motion synthesis: balancing data fidelity and stylistic authenticity. IEEE Trans. Artif. Intell. 4 (2), 183–197 (2023).
Wei, S., Zhang, L. & Li, Y. Beyond biomechanics: the semantic gap in computational martial arts analysis. ACM Trans. Intell. Syst. Technol. 13 (4), 56 (2022).
Johnson, T., Miller, S. & Williams, R. Foundations of cross-modal learning: bridging representational gaps in heterogeneous data domains. J. Mach. Learn. Res. 23 (118), 1–34 (2022).
Goodfellow, I. et al. Generative adversarial nets. In Advances in Neural Information Processing Systems 2672–2680. (2014).
Chen, L., Zhang, H. & Xiao, J. Cross-modal generative adversarial networks: architectures, applications, and challenges. IEEE Trans. Pattern Anal. Mach. Intell. 43 (12), 4234–4250 (2021).
Zhang, K. et al. Adversarial spatio-temporal learning for video deblurring. IEEE Trans. Image Process. 28 (1), 291–301 (2019).
Zhang, K., Li, D., Luo, W., Ren, W. & Liu, W. Enhanced spatio-temporal interaction learning for video deraining: faster and better. IEEE Trans. Pattern Anal. Mach. Intell. 45 (1), 1287–1293 (2023).
Zhang, K. et al. MC-Blur: a comprehensive benchmark for image deblurring. IEEE Trans. Circuits Syst. Video Technol. 33 (10), 5916–5930 (2023).
Xu, T., Zhang, P. & Huang, Q. AttnGAN and beyond: text-to-image synthesis with generative adversarial networks. In Computer Vision: A Reference Guide 1–23. Springer. (2022).
Li, W., Zhang, P. & Zhang, L. Progressive refinement strategies in cross-modal generative models: from coarse to fine-grained correspondences. Int. J. Comput. Vis.. 131 (4), 940–957 (2023).
Anderson, P., He, X. & Buehler, C. Vision-language models for multimodal understanding and generation. Found. Trends Mach. Learn. 14 (2), 201–308 (2021).
Lu, Y. et al. TransFlow: Transformer as flow learner. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) 18063–18073 (2023).
Yan, L. et al. GL-RG: Global-local representation granularity for video captioning. In Proceedings of the International Joint Conference on Artificial Intelligence (IJCAI) 2769–2775 (2022).
Wang, J. et al. Text is MASS: modeling as stochastic embedding for text-video retrieval. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) 16551–16560 (2024).
Wang, J., Chen, K. & Yu, R. Human motion modeling challenges: dimensionality, temporality, and physical constraints. IEEE Trans. Vis. Comput. Graph. 28 (7), 2471–2486 (2022).
Yang, Z., Zhao, J. & Dhingra, B. Transformers for human motion modeling: Architectures, applications, and future directions. ACM Comput. Surv. 55 (9), 1–35 (2023).
Tevet, G. et al. Human motion diffusion model. In Proceedings of the International Conference on Learning Representations (ICLR) (2023).
Zhang, J. et al. T2M-GPT: generating human motion from textual descriptions with discrete representations. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) 14730–14740 (2023).
Karunratanakul, M., Preechakul, K., Suwajanakorn, S. & Tang, S. Guided motion diffusion for controllable human motion synthesis. In: Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV) 2151–2162 (2023).
Liu, H., Chen, T. & Wang, Q. Semantic alignment in cross-modal generation: challenges and solutions. In Proceedings of the European Conference on Computer Vision 435–451. (2022).
Zhang, R., Yang, Y. & Li, W. Preserving structural coherence in high-dimensional temporal data generation: new frontiers and persistent challenges. Neural Comput. 35 (5), 891–913 (2023).
Wilson, M., Taylor, J. & Brown, A. Knowledge graphs for cultural heritage: semantic foundations for digital preservation. J. Doc. 77 (4), 959–980 (2021).
Chen, J., Wang, L. & Zhang, Y. Structured knowledge representation for intangible cultural heritage: case studies in traditional craftsmanship and performance arts. J. Cult. Herit. 54, 178–192 (2022).
Davis, R., Chen, X. & Wilson, T. Knowledge graph technologies for cultural preservation: enhanced query capabilities and Inferential reasoning. Digit. Scholarsh. Humanit. 38 (1), 43–62 (2023).
Wang, Y., Liu, J. & Zhang, X. Construction and application of knowledge graph for Chinese martial arts cultural heritage. Digit. Herit. Int. Congress 234–241 (2021).
Smith, K., Johnson, L. & Davis, R. Bridging the semantic gap in multimodal martial arts documentation through structured knowledge representation. In Proceedings of the International Conference on Knowledge Capture 217–226. (2023).
Wu, T., Li, J. & Chen, M. Cross-modal alignment with knowledge graphs: a survey of methods and applications. IEEE Trans. Knowl. Data Eng. 33 (11), 3534–3549 (2021).
Wang, Y., Zhang, X. & Liu, T. Knowledge-anchored multimodal correspondence learning in traditional martial arts documentation. Pattern Recognit. Lett. 156, 187–194 (2022).
Beyer, L. et al. AMD: automatic multi-step distillation of large-scale vision models. In Proceedings of the European Conference on Computer Vision (ECCV) 437–454 (2024).
Li, Q., Zhang, R. & Wang, J. Knowledge-guided motion synthesis: enhancing biomechanical plausibility through structured domain constraints. In Proceedings of the ACM International Conference on Multimedia 2731–2740. (2023).
Chen, Y., Wu, Z. & Li, T. Hierarchical entity taxonomy design for martial arts knowledge representation: balancing technical precision and conceptual expressivity. Knowl. Organ. 50 (2), 123–142 (2023).
Wang, L., Zhang, J. & Chen, K. Multi-source knowledge extraction for martial arts domain: combining textual analysis with expert elicitation. In Proceedings of the International Conference on Knowledge Engineering and Knowledge Management 328–343. (2022).
Zhao, T., Li, R. & Wang, Y. Relationship schema design for martial arts knowledge graphs: capturing hierarchical, compositional, and conceptual connections. J. Inf. Sci. 49 (3), 321–338 (2023).
Liu, M., Chen, W. & Zhang, H. Attribute parameterization in martial arts knowledge models: from qualitative principles to quantitative execution parameters. Inf. Process. Manag. 59 (2), 102762 (2022).
Schlichtkrull, M. et al. Modeling relational data with graph convolutional networks. In Proceedings of the European Semantic Web Conference 593–607. (2018).
Wang, Z., Chen, T. & Li, Y. Domain-specific constraints in graph representation learning: applications to martial arts knowledge modeling. In Proceedings of the ACM SIGKDD Conference on Knowledge Discovery and Data Mining 3475–3484. (2022).
Zhang, H., Li, W. & Liu, Y. Comparative evaluation of knowledge embedding approaches for martial arts relationship prediction. Knowl. Based Syst. 258, 110182 (2023).
Chen, J., Li, T. & Wang, R. Multi-level attention mechanisms for spatial-temporal feature extraction in martial arts video analysis. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition 7389–7398. (2022).
Wu, Y., Zhang, M. & Li, J. Domain-adapted transformers for martial arts textual understanding: capturing technical terminology and conceptual foundations. Nat. Lang. Eng. 29 (2), 245–267 (2023).
Yang, C., Liu, R. & Chen, L. Spatial-temporal graph convolutional networks for skeleton-based martial arts motion analysis. IEEE Trans. Neural Netw. Learn. Syst. 33 (7), 3118–3132 (2022).
Wang, T., Zhang, H. & Chen, Y. Knowledge-guided cross-modal alignment for martial arts movement analysis: preventing spurious correlations through domain constraints. Pattern Recognit. 136, 109171 (2023).
Li, W., Zhao, J. & Wang, L. Zero-shot and few-shot learning in cross-modal martial arts technique recognition: Leveraging knowledge graph relationships. In Proceedings of the European Conference on Computer Vision 289–305. (2022).
Zhang, M., Li, T. & Wang, Y. Transformer-based motion generation with knowledge-conditioned attention mechanisms. In Proceedings of the International Conference on Machine Learning 15782–15791. (2023).
Chen, K., Wang, Z. & Liu, J. Multi-level discrimination strategies for human motion evaluation: from joint-level assessment to holistic technique appraisal. IEEE Trans. Hum. Mach. Syst. 52 (3), 367–379 (2022).
Li, Y., Chen, J. & Zhang, L. Direct knowledge injection methodologies in deep generative frameworks: applications to martial arts motion synthesis. Neural Netw. 161, 368–382 (2023).
Wang, R., Liu, H. & Chen, T. Graph attention networks for dynamic knowledge subgraph processing in motion generation tasks. In Proceedings of the AAAI Conference on Artificial Intelligence 7842–7850. (2022).
Chen, Y., Zhang, W. & Li, M. Knowledge consistency constraints for physically and stylistically authentic martial arts motion generation. In Proceedings of the IEEE International Conference on Robotics and Automation 8754–8760. (2023).
Zhang, L., Wang, J. & Chen, R. Hierarchical phase modeling in traditional martial arts: decomposing complex techniques for enhanced Temporal constraint specification. Comput. Animat. Virtual Worlds, 33(3–4), e2015 (2022).
Liu, T., Chen, K. & Wang, Y. Knowledge-enhanced generative approaches for limited-data domains: advances in cultural heritage documentation. IEEE Trans. Pattern Anal. Mach. Intell. 45 (6), 6842–6857 (2023).
Wang, L., Zhang, T. & Chen, Y. Multimodal martial arts dataset construction: protocols, challenges, and quality control mechanisms. Data Brief. 42, 108293 (2022).
Chen, J., Wang, Z. & Li, T. Structured annotation methodologies for martial arts technique documentation: balancing procedural detail with conceptual depth. J. Doc. 79 (1), 97–115 (2023).
Li, W., Zhang, H. & Wang, Y. Semi-automated alignment between multimodal data and knowledge graph entities in martial arts documentation. Inf. Process. Manag. 59 (3), 102908 (2022).
Zhang, M., Chen, T. & Wang, L. Computational infrastructure for knowledge-enhanced generative modeling: hardware configurations and software ecosystems. J. Supercomput. 79 (5), 6123–6142 (2023).
Wang, Z., Li, Y. & Chen, J. Comprehensive evaluation metrics for motion reconstruction in cultural heritage applications. Multimed. Tools Appl. 81 (8), 11437–11456 (2022).
Chen, Y., Wu, Z. & Zhang, T. Optimization strategies for knowledge-enhanced generative adversarial networks in martial arts motion synthesis. J. Real-Time Image Proc. 20 (1), 28–43 (2023).
Ionescu, C., Papava, D., Olaru, V. & Sminchisescu, C. Human3.6 M: large scale datasets and predictive methods for 3D human sensing in natural environments. IEEE Trans. Pattern Anal. Mach. Intell. 36 (7), 1325–1339 (2014).
Wang, J., Chen, K. & Zhang, L. Knowledge-guided evaluation metrics for martial arts motion assessment: beyond kinematic accuracy. IEEE Trans. Cybern. 52 (11), 11784–11797 (2022).
Li, T., Wang, Y. & Chen, J. Knowledge-constrained regularization for complex martial arts technique reconstruction: addressing biomechanical ambiguity through domain principles. Pattern Recogn. 138, 109218 (2023).
Chen, R., Zhang, H. & Wang, L. Comparative analysis of internal and external martial arts styles: implications for motion capture and reconstruction methodologies. Sports Eng. 25 (1), 12 (2022).
Zhang, T., Li, W. & Wang, Z. Limitations in extreme motion capture: quantifying information loss in high-velocity martial arts movements. IEEE Sens. J. 23 (4), 3829–3842 (2023).
Chen, K., Li, W. & Zhang, M. Systematic ablation study methodologies for knowledge-integrated generative models: quantifying component contributions. In Proceedings of the International Conference on Learning Representations 1–15 (2023).
Zhang, H., Wang, Y. & Chen, J. The critical role of domain-specific knowledge in martial arts motion generation: evidence from comparative performance analysis. Neural Comput. 34 (8), 1785–1812 (2022).
Li, T., Chen, R. & Wang, Z. Cross-modal attention mechanisms for bridging visual appearances and biomechanical principles in martial arts technique analysis. IEEE Trans. Multimed. 25 (5), 2837–2851 (2023).
Wang, L., Zhang, H. & Li, W. Temporal constraint mechanisms for continuous versus segmented martial arts movements: style-specific considerations. In Proceedings of the ACM International Conference on Multimedia 2458–2467 (2022).
Chen, J., Wang, Y. & Zhang, T. Parameter sensitivity quantification methodologies for complex generative models: applications in cultural heritage preservation. J. Cult. Anal. 8 (2), 124–143 (2023).
Li, M., Chen, K. & Wang, L. Optimal loss balancing strategies for knowledge-enhanced motion generation in traditional martial arts. In Proceedings of the International Conference on Learning Representations 1–15 (2022).
Wang, Y., Li, T. & Chen, J. Knowledge integration strategies for preserving subtle technical nuances in martial arts motion reconstruction. IEEE Trans. Cult. Herit. 15 (2), 467–483 (2023).
Chen, K., Zhang, H. & Wang, L. Experimental validation of knowledge-enhanced generative models across diverse martial arts styles: technical accuracy and semantic fidelity assessment. Pattern Recogn. 146, 110097 (2024).
Li, W., Wang, Z. & Chen, Y. Domain-specific knowledge versus data-driven learning in martial arts motion generation: complementary strengths and integration approaches. Neural Netw. 159, 247–262 (2023).
Wang, Y., Zhang, T. & Chen, J. Current limitations in martial arts motion reconstruction: rapid movements, rare techniques, and knowledge representation challenges. Comput. Vis. Image Underst. 229, 103608 (2024).
Zhang, H., Li, T. & Wang, Y. Future directions in interactive martial arts learning systems: wearable sensing, augmented feedback, and adaptive guidance. IEEE Trans. Hum. Mach. Syst. 53 (1), 58–72 (2023).
Chen, J., Wang, Z. & Li, W. Cross-domain applications of knowledge-enhanced generative frameworks: from martial arts to diverse intangible cultural heritage preservation. Digit. Appl. Archaeol. Cult. Herit., 27, e00262 (2024).
Chen, X., Wang, Y. & Liu, T. Chinese martial arts (Wushu): historical development and contemporary cultural significance. J. Cult. Stud. 45 (2), 187–203 (2020).
Zhang, H., Li, W. & Chen, Y. Challenges and opportunities in the inheritance of traditional martial arts techniques: a systematic analysis. Int. J. Cult. Herit. 15 (4), 478–495 (2022).
Johnson, R. & Smith, K. Traditional knowledge transmission mechanisms in martial arts: limitations and digital transformation paths. J. Cult. Preserv. 12 (3), 267–285 (2021).
Wang, L., Chen, J. & Zhang, Q. Biomechanical analysis of traditional Chinese martial arts: energy flow patterns and technical execution principles. Sports Biomech. Int. 28 (4), 412–429 (2023).
Liu, Y., Zhao, J. & Wu, T. Motion capture technologies for martial arts documentation: a comparative analysis of optical, inertial, and vision-based approaches. IEEE Trans. Vis. Comput. Graph. 29 (2), 1237–1252 (2022).
Chen, H., Wang, Z. & Li, R. Interpretable movement analysis: bridging the gap between motion data and semantic understanding in martial arts. Artif. Intell. Rev. 40 (3), 567–588 (2021).
Wang, Q., Li, S. & Zhang, R. Knowledge graphs for complex domain representation: advances and applications. J. Knowl. Eng. 37 (2), 145–162 (2023).
Yang, M., Chen, X. & Zhao, T. Generative adversarial networks for cultural data synthesis: applications and ethical considerations. IEEE Trans. Neural Netw. Learn. Syst. 33 (9), 4128–4141 (2022).
Zhang, J., Li, H. & Wu, Y. Cross-modal learning frameworks for integrated Understanding of human movement: from perception to generation. Comput. Vis. Image Underst. 227, 103568 (2023).
Brown, M., Johnson, K. & Davis, L. Evolution of martial arts motion documentation: from manual annotation to AI-driven analysis. Int. J. Comput. Vis. 128 (6), 1534–1550 (2020).
Smith, A., Williams, J. & Jones, R. Marker-based motion capture systems for martial arts performance analysis: capabilities and constraints. J. Sports Sci. 39 (5), 518–532 (2021).
Chen, W., Liu, Y. & Zhang, H. Limitations of marker-based systems in capturing authentic martial arts movements: a quantitative assessment. IEEE Sens. J. 22 (7), 6921–6937 (2022).
Park, S., Kim, J. & Lee, M. Markerless vision-based human pose estimation for martial arts documentation: current status and challenges. Comput. Vis. Pattern Recognit. Ann. Rev. 15, 287–304 (2021).
Xu, H., Li, W. & Zhang, T. High-velocity movement tracking in markerless motion capture: performance analysis and optimization strategies. IEEE Trans. Pattern Anal. Mach. Intell. 44 (8), 4562–4578 (2022).
Garcia, J., Martinez, A. & Rodriguez, P. Addressing occlusion challenges in complex martial arts movements: multi-view fusion approaches. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition 3258–3267 (2023).
Wang, L., Chen, R. & Yang, Z. IMU-based motion capture for martial arts: system design and performance evaluation. Sens. Actuators A Phys. 321, 112587 (2021).
Zhang, K., Wu, Y. & Liu, J. Drift compensation in inertial measurement systems for long-duration martial arts performance capture. IEEE Trans. Instrum. Meas. 71, 1–12 (2022).
Li, R., Chen, T. & Wang, M. End-to-end deep learning frameworks for martial arts motion reconstruction: a comparative study. Neural Netw. 158, 174–188 (2023).
Zhao, J., Tao, D. & Wen, L. Transformer-based temporal modeling for complex human movement sequences in traditional martial arts. In Proceedings of the International Conference on Machine Learning 11824–11833 (2022).
Liu, Y., Wang, Z. & Chen, H. Challenges in AI-driven martial arts motion synthesis: balancing data fidelity and stylistic authenticity. IEEE Trans. Artif. Intell. 4 (2), 183–197 (2023).
Wei, S., Zhang, L. & Li, Y. Beyond biomechanics: the semantic gap in computational martial arts analysis. ACM Trans. Intell. Syst. Technol. 13 (4), 56 (2022).
Johnson, T., Miller, S. & Williams, R. Foundations of cross-modal learning: bridging representational gaps in heterogeneous data domains. J. Mach. Learn. Res. 23 (118), 1–34 (2022).
Goodfellow, I. et al. Generative adversarial nets. In Advances in Neural Information Processing Systems 2672–2680. (2014).
Chen, L., Zhang, H. & Xiao, J. Cross-modal generative adversarial networks: architectures, applications, and challenges. IEEE Trans. Pattern Anal. Mach. Intell. 43 (12), 4234–4250 (2021).
Zhang, K. et al. Adversarial spatio-temporal learning for video deblurring. IEEE Trans. Image Process. 28 (1), 291–301 (2019).
Zhang, K., Li, D., Luo, W., Ren, W. & Liu, W. Enhanced spatio-temporal interaction learning for video deraining: faster and better. IEEE Trans. Pattern Anal. Mach. Intell. 45 (1), 1287–1293 (2023).
Zhang, K. et al. MC-Blur: a comprehensive benchmark for image deblurring. IEEE Trans. Circuits Syst. Video Technol. 33 (10), 5916–5930 (2023).
Xu, T., Zhang, P. & Huang, Q. AttnGAN and beyond: text-to-image synthesis with generative adversarial networks. In Computer Vision: A Reference Guide 1–23. Springer. (2022).
Li, W., Zhang, P. & Zhang, L. Progressive refinement strategies in cross-modal generative models: from coarse to fine-grained correspondences. Int. J. Comput. Vis.. 131 (4), 940–957 (2023).
Anderson, P., He, X. & Buehler, C. Vision-language models for multimodal understanding and generation. Found. Trends Mach. Learn. 14 (2), 201–308 (2021).
Lu, Y. et al. TransFlow: Transformer as flow learner. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) 18063–18073 (2023).
Yan, L. et al. GL-RG: Global-local representation granularity for video captioning. In Proceedings of the International Joint Conference on Artificial Intelligence (IJCAI) 2769–2775 (2022).
Wang, J. et al. Text is MASS: modeling as stochastic embedding for text-video retrieval. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) 16551–16560 (2024).
Wang, J., Chen, K. & Yu, R. Human motion modeling challenges: dimensionality, temporality, and physical constraints. IEEE Trans. Vis. Comput. Graph. 28 (7), 2471–2486 (2022).
Yang, Z., Zhao, J. & Dhingra, B. Transformers for human motion modeling: Architectures, applications, and future directions. ACM Comput. Surv. 55 (9), 1–35 (2023).
Tevet, G. et al. Human motion diffusion model. In Proceedings of the International Conference on Learning Representations (ICLR) (2023).
Zhang, J. et al. T2M-GPT: generating human motion from textual descriptions with discrete representations. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) 14730–14740 (2023).
Karunratanakul, M., Preechakul, K., Suwajanakorn, S. & Tang, S. Guided motion diffusion for controllable human motion synthesis. In: Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV) 2151–2162 (2023).
Liu, H., Chen, T. & Wang, Q. Semantic alignment in cross-modal generation: challenges and solutions. In Proceedings of the European Conference on Computer Vision 435–451. (2022).
Zhang, R., Yang, Y. & Li, W. Preserving structural coherence in high-dimensional temporal data generation: new frontiers and persistent challenges. Neural Comput. 35 (5), 891–913 (2023).
Wilson, M., Taylor, J. & Brown, A. Knowledge graphs for cultural heritage: semantic foundations for digital preservation. J. Doc. 77 (4), 959–980 (2021).
Chen, J., Wang, L. & Zhang, Y. Structured knowledge representation for intangible cultural heritage: case studies in traditional craftsmanship and performance arts. J. Cult. Herit. 54, 178–192 (2022).
Davis, R., Chen, X. & Wilson, T. Knowledge graph technologies for cultural preservation: enhanced query capabilities and Inferential reasoning. Digit. Scholarsh. Humanit. 38 (1), 43–62 (2023).
Wang, Y., Liu, J. & Zhang, X. Construction and application of knowledge graph for Chinese martial arts cultural heritage. Digit. Herit. Int. Congress 234–241 (2021).
Smith, K., Johnson, L. & Davis, R. Bridging the semantic gap in multimodal martial arts documentation through structured knowledge representation. In Proceedings of the International Conference on Knowledge Capture 217–226. (2023).
Wu, T., Li, J. & Chen, M. Cross-modal alignment with knowledge graphs: a survey of methods and applications. IEEE Trans. Knowl. Data Eng. 33 (11), 3534–3549 (2021).
Wang, Y., Zhang, X. & Liu, T. Knowledge-anchored multimodal correspondence learning in traditional martial arts documentation. Pattern Recognit. Lett. 156, 187–194 (2022).
Beyer, L. et al. AMD: automatic multi-step distillation of large-scale vision models. In Proceedings of the European Conference on Computer Vision (ECCV) 437–454 (2024).
Li, Q., Zhang, R. & Wang, J. Knowledge-guided motion synthesis: enhancing biomechanical plausibility through structured domain constraints. In Proceedings of the ACM International Conference on Multimedia 2731–2740. (2023).
Chen, Y., Wu, Z. & Li, T. Hierarchical entity taxonomy design for martial arts knowledge representation: balancing technical precision and conceptual expressivity. Knowl. Organ. 50 (2), 123–142 (2023).
Wang, L., Zhang, J. & Chen, K. Multi-source knowledge extraction for martial arts domain: combining textual analysis with expert elicitation. In Proceedings of the International Conference on Knowledge Engineering and Knowledge Management 328–343. (2022).
Zhao, T., Li, R. & Wang, Y. Relationship schema design for martial arts knowledge graphs: capturing hierarchical, compositional, and conceptual connections. J. Inf. Sci. 49 (3), 321–338 (2023).
Liu, M., Chen, W. & Zhang, H. Attribute parameterization in martial arts knowledge models: from qualitative principles to quantitative execution parameters. Inf. Process. Manag. 59 (2), 102762 (2022).
Schlichtkrull, M. et al. Modeling relational data with graph convolutional networks. In Proceedings of the European Semantic Web Conference 593–607. (2018).
Wang, Z., Chen, T. & Li, Y. Domain-specific constraints in graph representation learning: applications to martial arts knowledge modeling. In Proceedings of the ACM SIGKDD Conference on Knowledge Discovery and Data Mining 3475–3484. (2022).
Zhang, H., Li, W. & Liu, Y. Comparative evaluation of knowledge embedding approaches for martial arts relationship prediction. Knowl. Based Syst. 258, 110182 (2023).
Chen, J., Li, T. & Wang, R. Multi-level attention mechanisms for spatial-temporal feature extraction in martial arts video analysis. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition 7389–7398. (2022).
Wu, Y., Zhang, M. & Li, J. Domain-adapted transformers for martial arts textual understanding: capturing technical terminology and conceptual foundations. Nat. Lang. Eng. 29 (2), 245–267 (2023).
Yang, C., Liu, R. & Chen, L. Spatial-temporal graph convolutional networks for skeleton-based martial arts motion analysis. IEEE Trans. Neural Netw. Learn. Syst. 33 (7), 3118–3132 (2022).
Wang, T., Zhang, H. & Chen, Y. Knowledge-guided cross-modal alignment for martial arts movement analysis: preventing spurious correlations through domain constraints. Pattern Recognit. 136, 109171 (2023).
Li, W., Zhao, J. & Wang, L. Zero-shot and few-shot learning in cross-modal martial arts technique recognition: Leveraging knowledge graph relationships. In Proceedings of the European Conference on Computer Vision 289–305. (2022).
Zhang, M., Li, T. & Wang, Y. Transformer-based motion generation with knowledge-conditioned attention mechanisms. In Proceedings of the International Conference on Machine Learning 15782–15791. (2023).
Chen, K., Wang, Z. & Liu, J. Multi-level discrimination strategies for human motion evaluation: from joint-level assessment to holistic technique appraisal. IEEE Trans. Hum. Mach. Syst. 52 (3), 367–379 (2022).
Li, Y., Chen, J. & Zhang, L. Direct knowledge injection methodologies in deep generative frameworks: applications to martial arts motion synthesis. Neural Netw. 161, 368–382 (2023).
Wang, R., Liu, H. & Chen, T. Graph attention networks for dynamic knowledge subgraph processing in motion generation tasks. In Proceedings of the AAAI Conference on Artificial Intelligence 7842–7850. (2022).
Chen, Y., Zhang, W. & Li, M. Knowledge consistency constraints for physically and stylistically authentic martial arts motion generation. In Proceedings of the IEEE International Conference on Robotics and Automation 8754–8760. (2023).
Zhang, L., Wang, J. & Chen, R. Hierarchical phase modeling in traditional martial arts: decomposing complex techniques for enhanced Temporal constraint specification. Comput. Animat. Virtual Worlds, 33(3–4), e2015 (2022).
Liu, T., Chen, K. & Wang, Y. Knowledge-enhanced generative approaches for limited-data domains: advances in cultural heritage documentation. IEEE Trans. Pattern Anal. Mach. Intell. 45 (6), 6842–6857 (2023).
Wang, L., Zhang, T. & Chen, Y. Multimodal martial arts dataset construction: protocols, challenges, and quality control mechanisms. Data Brief. 42, 108293 (2022).
Chen, J., Wang, Z. & Li, T. Structured annotation methodologies for martial arts technique documentation: balancing procedural detail with conceptual depth. J. Doc. 79 (1), 97–115 (2023).
Li, W., Zhang, H. & Wang, Y. Semi-automated alignment between multimodal data and knowledge graph entities in martial arts documentation. Inf. Process. Manag. 59 (3), 102908 (2022).
Zhang, M., Chen, T. & Wang, L. Computational infrastructure for knowledge-enhanced generative modeling: hardware configurations and software ecosystems. J. Supercomput. 79 (5), 6123–6142 (2023).
Wang, Z., Li, Y. & Chen, J. Comprehensive evaluation metrics for motion reconstruction in cultural heritage applications. Multimed. Tools Appl. 81 (8), 11437–11456 (2022).
Chen, Y., Wu, Z. & Zhang, T. Optimization strategies for knowledge-enhanced generative adversarial networks in martial arts motion synthesis. J. Real-Time Image Proc. 20 (1), 28–43 (2023).
Ionescu, C., Papava, D., Olaru, V. & Sminchisescu, C. Human3.6 M: large scale datasets and predictive methods for 3D human sensing in natural environments. IEEE Trans. Pattern Anal. Mach. Intell. 36 (7), 1325–1339 (2014).
Wang, J., Chen, K. & Zhang, L. Knowledge-guided evaluation metrics for martial arts motion assessment: beyond kinematic accuracy. IEEE Trans. Cybern. 52 (11), 11784–11797 (2022).
Li, T., Wang, Y. & Chen, J. Knowledge-constrained regularization for complex martial arts technique reconstruction: addressing biomechanical ambiguity through domain principles. Pattern Recogn. 138, 109218 (2023).
Chen, R., Zhang, H. & Wang, L. Comparative analysis of internal and external martial arts styles: implications for motion capture and reconstruction methodologies. Sports Eng. 25 (1), 12 (2022).
Zhang, T., Li, W. & Wang, Z. Limitations in extreme motion capture: quantifying information loss in high-velocity martial arts movements. IEEE Sens. J. 23 (4), 3829–3842 (2023).
Chen, K., Li, W. & Zhang, M. Systematic ablation study methodologies for knowledge-integrated generative models: quantifying component contributions. In Proceedings of the International Conference on Learning Representations 1–15 (2023).
Zhang, H., Wang, Y. & Chen, J. The critical role of domain-specific knowledge in martial arts motion generation: evidence from comparative performance analysis. Neural Comput. 34 (8), 1785–1812 (2022).
Li, T., Chen, R. & Wang, Z. Cross-modal attention mechanisms for bridging visual appearances and biomechanical principles in martial arts technique analysis. IEEE Trans. Multimed. 25 (5), 2837–2851 (2023).
Wang, L., Zhang, H. & Li, W. Temporal constraint mechanisms for continuous versus segmented martial arts movements: style-specific considerations. In Proceedings of the ACM International Conference on Multimedia 2458–2467 (2022).
Chen, J., Wang, Y. & Zhang, T. Parameter sensitivity quantification methodologies for complex generative models: applications in cultural heritage preservation. J. Cult. Anal. 8 (2), 124–143 (2023).
Li, M., Chen, K. & Wang, L. Optimal loss balancing strategies for knowledge-enhanced motion generation in traditional martial arts. In Proceedings of the International Conference on Learning Representations 1–15 (2022).
Wang, Y., Li, T. & Chen, J. Knowledge integration strategies for preserving subtle technical nuances in martial arts motion reconstruction. IEEE Trans. Cult. Herit. 15 (2), 467–483 (2023).
Chen, K., Zhang, H. & Wang, L. Experimental validation of knowledge-enhanced generative models across diverse martial arts styles: technical accuracy and semantic fidelity assessment. Pattern Recogn. 146, 110097 (2024).
Li, W., Wang, Z. & Chen, Y. Domain-specific knowledge versus data-driven learning in martial arts motion generation: complementary strengths and integration approaches. Neural Netw. 159, 247–262 (2023).
Wang, Y., Zhang, T. & Chen, J. Current limitations in martial arts motion reconstruction: rapid movements, rare techniques, and knowledge representation challenges. Comput. Vis. Image Underst. 229, 103608 (2024).
Zhang, H., Li, T. & Wang, Y. Future directions in interactive martial arts learning systems: wearable sensing, augmented feedback, and adaptive guidance. IEEE Trans. Hum. Mach. Syst. 53 (1), 58–72 (2023).
Chen, J., Wang, Z. & Li, W. Cross-domain applications of knowledge-enhanced generative frameworks: from martial arts to diverse intangible cultural heritage preservation. Digit. Appl. Archaeol. Cult. Herit., 27, e00262 (2024).
Contributed Indexing:
Keywords: Cross-modal learning; Cultural heritage preservation; Generative adversarial networks; Knowledge graph; Martial arts; Motion reconstruction
Entry Date(s):
Date Created: 20260121 Latest Revision: 20260121
Update Code:
20260122
DOI:
10.1038/s41598-026-36095-z
PMID:
41565884
Database:
MEDLINE
Weitere Informationen
This paper presents a novel knowledge graph enhanced cross-modal generative adversarial network (KG-CMGAN) for preserving traditional martial arts techniques. We address the challenges of capturing the complex, multidimensional nature of martial arts by integrating structured domain knowledge with advanced deep learning architectures. Our framework establishes an end-to-end solution that bridges visual, textual, and sequential representations to achieve comprehensive motion reconstruction while preserving stylistic authenticity and semantic meaning. The proposed approach includes a comprehensive martial arts knowledge graph that formalizes domain-specific ontology, a knowledge-guided cross-modal alignment mechanism that effectively integrates heterogeneous data sources, and a knowledge-enhanced adversarial learning architecture specifically optimized for martial arts motion reconstruction. Extensive experiments across six traditional Chinese martial arts styles demonstrate significant improvements over state-of-the-art baselines, with 28.4% reduction in joint position error and 91.2% knowledge consistency score. Ablation studies confirm that knowledge graph integration is critical for generating culturally authentic movements. This research contributes a novel methodology for intangible cultural heritage preservation that captures both the physical execution and conceptual foundations of traditional martial arts.
(© 2026. The Author(s).)
Declarations. Competing interests: The authors declare no competing interests.