Digital Art Design and Intelligent Re-creation of Intangible Cultural Heritage Patterns Based on Diffusion Generation Model

With the increasing demand for the protection and dissemination of intangible cultural heritage (ICH), the digitization and innovative redesign of traditional patterns have become important directions for cultural inheritance. This paper proposes a digital art design and intelligent re-creation method for ICH patterns based on the Diffusion Model (DM). First, a database containing 4320 high-resolution ICH patterns is constructed, and semantic correspondence is achieved through feature annotation. Second, an improved Latent Diffusion Model (LDM) is adopted, introducing an attention-guided module and a cultural semantic embedding layer to enhance the model's understanding of the semantic and stylistic features of the patterns. Then, through style transfer and feature fusion algorithms, the diverse generation and innovative re-creation of ICH pattern styles are achieved. Finally, the effectiveness is evaluated using subjective aesthetic evaluation and the Structural Similarity Index (SSIM). The generated patterns achieve a structural fidelity of 0.956, a 12 % improvement in style consistency, and an average user aesthetic score of 1.4 . This study not only verified the feasibility and innovation of the diffusion model in the digital art design of intangible cultural heritage, but also provided a technical path with practical value for the intelligent transmission of cultural heritage.