AIGC-Driven Digital Fashion Performance Content Creation and Multi-Modal Communication Path Exploration

To address the inconsistency between the physical properties of eco-friendly textile materials and their visual representation in digital fashion generation, as well as the limited adaptability of multimodal communication under dynamic performance scenarios, this study proposes an Artificial Intelligence Generated Content (AIGC) framework integrating a Physics-Informed Neural Network (PINN) with a diffusion-based generation model and multimodal recommendation strategy. Material microstructure information and physical constraint equations are embedded into the generation process to jointly model permeability, glossiness, and elastic behavior, while motion-aware texture evolution and style-guided optimization improve the realism and physical consistency of virtual garments. A crossmedia adaptation mechanism further enables efficient deployment across AR fitting, virtual performances, social media, and metaverse applications through dynamic parameter mapping and low-latency transmission. Experimental results demonstrate superior recommendation accuracy, visual consistency, and computational efficiency, achieving an 89.7% Top-5 accuracy, an 83.2% Recall@10, and an 85 ms generation latency while maintaining stable physical perception of material properties. By bridging physically constrained content generation with multimodal communication, the proposed framework provides a reliable solution for intelligent digital fashion and sustainable textile visualization. Furthermore, its integration of real-time transmission and multimodal perception offers valuable insights for electromagnetic-enabled wireless media delivery, antenna-assisted immersive display systems, and future smart communication platforms requiring high-fidelity visual information propagation.