Jianing Luo and Boyuan Yu's project addresses the visual discord created by Hong Kong's rapid urbanization, proposing an innovative design methodology that employs Pix2Pix GAN to generate building facades in context with historical architecture. By compiling a dataset of 160 facade images from the city's old neighborhoods and processing them into a usable format, they trained the model to create designs that reflect local vernacular styles, showcasing high fidelity to existing forms while offering cleaner, stylized alternatives for early-stage architectural design.
The research culminated in the development of a methodology that effectively harmonizes new constructions with Hong Kong's diverse architectural landscape, utilizing AI for facade generation.
By leveraging Pix2Pix GAN, the project establishes a new design paradigm to ensure that modern architecture resonates with the historical context of Hong Kong.
Collection
[
|
...
]