Hehan Zhou
Home    Research    Projects    About



Here is a list of research  work that Hehan Zhou has been engaged in over the years, including her master's research at South China University of Technology (SCUT), academic collaborations, international conference papers, and forthcoming publications.

A multi-input CNN model that integrates terrain and blue-green spatial features to rapidly generate compound flood depth maps at the block scale, enhancing prediction accuracy and resilience evaluation
 
UPF-GAN generates high-accuracy flood depth maps from urban form images, running 640× faster than hydrodynamic models. It enables fast, interpretable resilience testing in dense urban blocks
 
Using Pix2Pix, this study generates street plans inspired by Great Streets, translating road inputs into walkable, green, and legible layouts with early signs of design logic


By combining Stable Diffusion and LoRA, this study explores AI-driven methods for post-war building restoration, bridging memory, resilience, and digital repair futures
 
Using heat and wind as core flows, this project trains a generative model to explore reciprocal dynamics between urban form and flow, guiding responsive spatial design
 
Linking MIKE21 simulations with SHAP analysis, this study uncovers how urban form drives flood resilience, offering interpretable insights for adaptive spatial design

Applying UPF-GAN to Weiyuan Island, this study maps flood resilience under P50/P100 scenarios and reveals form-risk relationships for real-time urban design evaluation
  
This study synthesizes resilience theory, morphological metrics, and AI techniques to build a conceptual foundation for integrated design-evaluation models like UPF-GAN