Preliminary Study on Architectural Skin Design Method Driven by Neural Style Transfer

With the development of new technologies and materials, architectural skins have become more complex and diverse, requiring a lot of effort to model different skins in architectural design. Artificial intelligence technology has a deep impact on different disciplines and can improve the efficiency of architectural skin design and inspire architects. In order to find new expressions of the building skin, it is necessary to focus on the dialogue between the architectural skin and the environment, and to innovate by breaking away from the patterned and formalized design. This work combines the development of artificial intelligence to initially explore the role of neural style transfer for architectural skin design, where the fusion of different styles of architectural images is programmed to generate another style, thus allowing architects to make clearer decisions and judgments. After performing neural style migration on five different types of buildings, the results obtained can present the new style form more completely, which provides new ideas and inspiration for the application of artificial intelligence in architectural design.