Revealing the pathways among visual exposures, place perceptions, and mental health considering individual mobility

Abstract

Using individual mobility and street view image data collected in Hong Kong, this study seeks to advance our understanding of the impacts of people’s dynamic visual exposures on their mental health by considering the indirect effects of their subjective perceptions of places they visit during their daily mobility. First, we extracted a range of street view features (SVFs) of visual exposures and place perceptions from extensive street view images using deep learning models. Second, individuals’ mobility-based visual exposures and place perceptions were estimated by linking these SVFs to their daily movement trajectories. Further, a causal inference model was employed to examine the direct and indirect pathways among individual visual exposures, place perceptions, and mental health. The findings reveal that (a) individual visual exposures significantly influence both place perceptions and mental health; (b) people’s place perceptions have mediating effects on exposure-health relationships; (c) traditional residence-based approaches tend to underestimate the health impacts of visual exposures and place perceptions.