An Interactive Application Study of AI Image Generation System Based on Stable Diffusion Under the Visual Logic of Surrealism
Haiyu Yuan, Yuebin Liao, Yu Shang · 2025
In the context of China's active promotion of the integration between artificial intelligence and cultural innovation, generative AI models are increasingly applied in the field of art and design. This study presents an image generation and interaction system that merges surrealist art concepts, constructed using Stable Diffusion and TouchDesigner. The system generates imaginative visual outcomes in real time through the collaboration of facial capture and text prompts, operating within a latent space. By analyzing the alignment between the diffusion mechanism and surrealist concepts such as “automatism” and the “unconscious,” this paper explores the creative potential of AI-generated imagery in artistic expression. Unlike previous experiments that aimed solely at image output, the system developed here emphasizes human-machine collaboration, visual interaction, and nonlinear generative processes. It offers practical technical solutions and creative paradigms for real-world applications in education, exhibitions, and cultural product design. The study demonstrates that the system not only ensures stable image output but also serves as an interactive platform that stimulates creativity and reconfigures aesthetic experience, thereby opening new ways for the practical application of AI in visual art.