AI-driven style transfer for enhancing cultural experience in Yangliuqing new year prints

Intangible cultural heritage, such as Yangliuqing New Year Prints, faces persistent challenges in engaging contemporary audiences within digital environments. This study presents ChineseNYP, an AI-mediated interactive system that combines on-demand facial stylization with real-time facial tracking and visualization. Stable Diffusion XL, a custom-trained Low-Rank Adaptation model, and Lineart ControlNet are used to generate Yangliuqing New Year Print-style portraits, while MediaPipe and TouchDesigner support responsive interaction with the generated images. The system was evaluated through an exploratory case-level technical comparison, assessments by four experts, a User Experience Questionnaire completed by 34 participants, and PLS-SEM analysis involving 127 participants. The expert assessments and user-experience results were generally positive. The PLS-SEM results showed significant positive associations among technological interaction, technology acceptance, embodied experience, and perceived cultural identity. These findings suggest that AI-mediated facial representation and real-time interaction may support engaging short-term experiences of traditional visual culture, while their longer-term implications for cultural dissemination, heritage preservation, and sustainability require further investigation.