KG-GAN: Knowledge Graph-Constrained GAN for Culturally Faithful Virtual Scene Synthesis
Abstract
This paper presents a knowledge-guided generative adversarial network (KG-GAN) for culturally faithful virtual scene construction and immersive experience optimization. A domain knowledge graph (KG) encodes traditional cultural entities and relations as graph embeddings, which condition the generator together with textual prompts. The semantic consistency discriminator calculates text-image cosine similarity based on CLIP cross-modal embedding, jointly scoring lexical semantic consistency and symbol reconstruction accuracy. The generator employs AdaIN for knowledge-conditional modulation, and multi-scale training and PWC-Net optical flow regularization collaboratively optimize local details and inter-frame stability. To balance global realism and local cultural details., we adopt a multi-scale training schedule and an attention controller that dynamically allocates textures and color palettes to culture-critical regions. Temporal stability is further improved with a lightweight consistency loss.Experiments on a curated cultural corpus and human-corrected pairs report cultural semantic fidelity of 89.7%–92.1%, relation compliance of 86.9%–88.6%, inter-frame SSIM under fast transitions of 0.782, and optical-flow smoothness of 0.751. A vocational Chinese-language classroom case study shows higher cultural expressiveness, learner engagement, and motion comfort compared with diffusion-based and CLIP-guided baselines, with ablations confirming the contributions of KG constraints, the semantic discriminator, and attention control. The framework is plug-and-play for XR/education/museums, and we provide KG schemas, code, and evaluation scripts to support reproducible deployment.References
Song L. Construction of socialist core values from the perspective of Chinese traditional culture[J]. International Journal of Frontiers in Sociology, 2021, 3(12): 69–76.
https://doi.org/10.25236/ijfs.2021.031210
Jiang B. Research on the application of Chinese traditional culture teaching in higher vocational education[J]. Kuram ve Uygulamada Egitim Bilimleri, 2022, 22(2): 1–13. (Result score too low)
Barrett A, Pack A, Guo Y, et al. Technology acceptance model and multi-user virtual reality learning environments for Chinese language education[J]. Interactive Learning Environments, 2023, 31(3): 1665–1682.
https://doi.org/10.1080/10494820.2020.1855209
Huang W. Digital realization of traditional residential houses in Fujian: application and research on immersive scenes in virtual animation design[J]. Procedia of Multidisciplinary Research, 2023, 1(9): 15–15. (Result score too low)
Sun H, Fan M, Sharma A. RETRACTED: Design and implementation of construction prediction and management platform based on building information modeling and three-dimensional simulation technology in Industry 4.0[J]. IET Collaborative Intelligent Manufacturing, 2021, 3(3): 224–232.
https://doi.org/10.1049/cim2.12019
Jin S, Fan M, Kadir A. Immersive spring morning in the Han Palace: learning traditional Chinese art via virtual reality and multi-touch tabletop[J]. International Journal of Human–Computer Interaction, 2022, 38(3): 213–226.
https://doi.org/10.1080/10447318.2021.1930389
Shadiev R, Wang X, Huang Y M. Cross-cultural learning in virtual reality environment: facilitating cross-cultural understanding, trait emotional intelligence, and sense of presence[J]. Educational Technology Research and Development, 2021, 69(5): 2917–2936.
https://doi.org/10.1007/s11423-021-10044-1
Wang Y, Wang L, Siau K L. Human-centered interaction in virtual worlds: a new era of generative artificial intelligence and metaverse[J]. International Journal of Human–Computer Interaction, 2025, 41(2): 1459–1501.
https://doi.org/10.1080/10447318.2024.2316376
Liu Q, Chen H, Crabbe M. Interactive study of multimedia and virtual technology in art education[J]. International Journal of Emerging Technologies in Learning (iJET), 2021, 16(1): 80–93.
https://doi.org/10.3991/ijet.v16i01.18227
Li W, Huang H, Solomon T, et al. Synthesizing personalized construction safety training scenarios for VR training[J]. IEEE Transactions on Visualization and Computer Graphics, 2022, 28(5): 1993–2002.
https://doi.org/10.1109/tvcg.2022.3150510
Qian J. Research on artificial intelligence technology of virtual reality teaching method in digital media art creation[J]. Journal of Internet Technology, 2022, 23(1): 125–132.
https://doi.org/10.53106/160792642022012301013
Pirker J, Dengel A. The potential of 360 virtual reality videos and real VR for education—a literature review[J]. IEEE Computer Graphics and Applications, 2021, 41(4): 76–89.
https://doi.org/10.1109/mcg.2021.3067999
Hammady R, Ma M, Al-Kalha Z, et al. A framework for constructing and evaluating the role of MR as a holographic virtual guide in museums[J]. Virtual Reality, 2021, 25(4): 895–918.
https://doi.org/10.1007/s10055-020-00497-9
Zhong H, Wang L, Zhang H. The application of virtual reality technology in the digital preservation of cultural heritage[J]. Computer Science and Information Systems, 2021, 18(2): 535–551.
https://doi.org/10.2298/csis200208009z
Karuzaki E, Partarakis N, Patsiouras N, et al. Realistic virtual humans for cultural heritage applications[J]. Heritage, 2021, 4(4): 4148–4171.
https://doi.org/10.3390/heritage4040228
Dong X, Mao Z, Sun Y, et al. Short-term wind power scenario generation based on conditional latent diffusion models[J]. IEEE Transactions on Sustainable Energy, 2023, 15(2): 1074–1085.
https://doi.org/10.1109/tste.2023.3327497
Peng J, Qiu R L J, Wynne J F, et al. CBCT‐based synthetic CT image generation using conditional denoising diffusion probabilistic model[J]. Medical Physics, 2024, 51(3): 1847–1859.
https://doi.org/10.1002/mp.16704
Zheng L. Artistic style image migration model based on cycle-consistent generative adversarial networks[J]. International Journal of Information and Communication Technology, 2025, 26(16): 53–68.
https://doi.org/10.1504/ijict.2025.146377
Yuan R, Wang B, Sun Y, et al. Conditional style-based generative adversarial networks for renewable scenario generation[J]. IEEE Transactions on Power Systems, 2022, 38(2): 1281–1296.
https://doi.org/10.1109/tpwrs.2022.3170992
Kang M, Zhu R, Chen D, et al. A cross-modal generative adversarial network for scenarios generation of renewable energy[J]. IEEE Transactions on Power Systems, 2023, 39(2): 2630–2640.
https://doi.org/10.1109/tpwrs.2023.3277698
Gan M, Wang C. Esophageal optical coherence tomography image synthesis using an adversarially learned variational autoencoder[J]. Biomedical Optics Express, 2022, 13(3): 1188–1201.
https://doi.org/10.1364/boe.449796
Wang Y, Ma X, Wang J, et al. Robust AUV visual loop-closure detection based on variational autoencoder network[J]. IEEE Transactions on Industrial Informatics, 2022, 18(12): 8829–8838.
https://doi.org/10.1109/tii.2022.3145860
Zhang P, Wang C, Kumar N, et al. Dynamic virtual network embedding algorithm based on graph convolution neural network and reinforcement learning[J]. IEEE Internet of Things Journal, 2021, 9(12): 9389–9398.
https://doi.org/10.1109/jiot.2021.3095094
Suárez-Varela J, Almasan P, Ferriol-Galmés M, et al. Graph neural networks for communication networks: context, use cases and opportunities[J]. IEEE Network, 2022, 37(3): 146–153.
https://doi.org/10.1109/mnet.123.2100773
Lim S H, Park J W. Hierarchical control strategy for effective virtual frequency responses of multiple WPPs[J]. IEEE Transactions on Power Systems, 2023, 39(1): 576–586.
https://doi.org/10.1109/tpwrs.2023.3269052
Carrasco J V, Ruiz F J, Soriano M C L. Empirical analysis of derived hierarchical responding[J]. International Journal of Psychology and Psychological Therapy, 2024, 24(2): 193–219. (Result score too low)
Wang J, Wu Q M J, Pourpanah F. An attentive-based generative model for medical image synthesis[J]. International Journal of Machine Learning and Cybernetics, 2023, 14(11): 3897–3910.
https://doi.org/10.1007/s13042-023-01871-0
Liu X, Pan J, Li X, et al. Attention based cross-domain synthesis and segmentation from unpaired medical images[J]. IEEE Transactions on Emerging Topics in Computational Intelligence, 2023, 8(1): 917–929.
https://doi.org/10.1109/tetci.2023.3296499
DOI:
https://doi.org/10.31449/inf.v50i10.12258Downloads
Published
How to Cite
Issue
Section
License
Authors retain copyright in their work. By submitting to and publishing with Informatica, authors grant the publisher (Slovene Society Informatika) the non-exclusive right to publish, reproduce, and distribute the article and to identify itself as the original publisher.
All articles are published under the Creative Commons Attribution license CC BY 3.0. Under this license, others may share and adapt the work for any purpose, provided appropriate credit is given and changes (if any) are indicated.
Authors may deposit and share the submitted version, accepted manuscript, and published version, provided the original publication in Informatica is properly cited.
