Semantic Understanding and Generation Mechanism of Multimodal Knowledge Graphs in Digital Media Art Creation
Year: 2026
Type: conference-paper
Abstract: This paper deeply studies the application of multimodal knowledge graphs in digital media art creation and proposes a multimodal fusion attention propagation algorithm (MFAP). The algorithm combines key technologies such as multimodal data preprocessing, semantic network construction and understanding, and creative element generation. Through experimental simulation on a large digital media art work dataset, this paper comprehensively evaluates the performance of the MFAP algorithm in semantic understanding and generation tasks. The results show that the MFAP algorithm significantly outperforms the multimodal Transformer and Diffusion models in key indicators such as accuracy, recall, and F1 value of semantic understanding. At the same time, the MFAP algorithm also shows a higher level in creative evaluation and user satisfaction of generated image, audio, and text works. This study not only provides a new technical path for digital media art creation, but also lays a solid foundation for the research on multimodal data fusion and generation mechanisms, which is of great significance to promoting innovation and development in the field of digital media art.
Authors: Peng Cui, Feng Chen, Siyang Liu, Qifan Zheng
Institutions: Dalian Art College, Luxun Academy of Fine Arts
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W7201868652