Reimagining tourism pedestrian street aesthetics through machine learning:Understanding the role of spatial form based on a case study in Luoyang,ChinaOA

Pedestrian streets are vital for urban livability,tourism,and cultural identity.This research examines how human-perspective spatial form influences aesthetic perception,using a tourist street in central Luoyang as a case site.Based on perceptual evaluations from participants in Luoyang and Xi''an,the research isolates key structural elements and reveals the underlying relationship between spatial form and tourist aesthetic preferences.Deep learning models were used to extract spatial indicators from real-world streetscapes,aligning them with abstracted representations.Modelling the extracted indicators with a Generalized Additive Model(GAM),the study enables large-scale analysis and captures both individual spatial characteristics and their interactive effects on aesthetic perception.This approach not only models complex nonlinear relationships but also provides a solid foundation for aesthetic prediction and assessment.The findings identify the proportion of sky(PS),ground area(PG),and spatial depth(D)as key factors influencing aesthetic judgments,while the proportion of vertical elements(PV)and the ground-to-vertical ratio(G/V)show high multicollinearity.Additionally,street-level average aesthetics tend to be rated higher than point-wise average aesthetics.These insights allow for the layout and adjustment of spatial form by balancing the aesthetic preferences of local and non-local visitors,ultimately enhancing pedestrian street aesthetics.