Publication

Our ECCV 2026 paper introduces a benchmark and multimodal framework for mapping gated communities and studying their equity implications.

Our Beyond fidelity study examines how levels of detail in virtual city models affect gaze behavior and perceived realism.

Our UST-GNN paper models spatial and network relationships in urban health prediction.

Our PIG-GNN study combines spatial interaction theory and graph learning to predict urban commuting flows.

Our Sustainable Cities and Society paper combines aerial and street-level perspectives to study green space during urban expansion.

Our Cities paper examines how visual variation along streets helps explain urban perception.

Our Cities paper formalises urban pattern language as a multi-scalar framework for understanding urban form.