Dutch Chinatown Façade Style Migration revitalizes Chinatown’s architectural landscape by transferring traditional Chinese design motifs onto Dutch building façades. Users work within a browser interface, where they assign RGB‑coded labels to elements such as roof, window, door, and column on an input canvas. Upon hitting a transfer button, the system generates a migrated façade. This lightweight approach not only engages non‑experts in cultural preservation but also introduces playful exploration of urban heritage through simple labeling.

The pipeline operates by merging a JavaScript‑driven front end, dynamic 3D canvas reconstruction, and a pix2pix conditional GAN trained on paired Chinese and Dutch façade datasets. After validation by an integrated discriminator, each generated output is rendered in a Three.js viewer for manual perspective refinement. Through the combination of web‑based labeling, generative AI, and interactive 3D visualization, the project establishes a novel methodology for participatory style migration and community‑driven heritage design.