Measuring the effectiveness of street renewal design: Insights from visual preference surveys, deep-learning technology, and eye-tracking simulation software

Abstract

This study investigates the impact of urban renewal projects (URPs) on the visual quality of street space (VQoS) through the evaluation of both psychological and visual perception, addressing the significant gap in empirical evidence. Analysing fifty-one URPs in densely populated Asian cities with complex urban settings, it assesses how measurable changes in spatial features such as greenness, walkability, and openness affect perceived VQoS pre- and post-renewal to identify the effective street improvements. The methodology includes capturing multi-year Google Street View images (SVIs) of selected URPs and conducting online visual preference surveys to gather data on perceived VQoS. Deep learning algorithms were used to identify key elements within the street scene. Additionally, eye-tracking emulation technology was utilized to investigate how individuals subconsciously interact with visual stimuli, aiming to determine whether variations in perceived VQoS are linked to how readily transformed street elements capture visual attention. Results indicate that VQoS improvements are predominantly associated with increased greenery, suggesting that enhanced green spaces contribute to higher visual appeal. There is also an unconscious preference for visually engaging with street elements, such as buildings and greenery, while less detailed areas, including sidewalks, driveways, and unadorned building façades, are often overlooked. Moreover, attention-catching designs of new building facades can significantly enhance VQoS. These findings suggest that urban design interventions should prioritize incorporating greenery and visually stimulating elements to create more engaging and visually appealing street spaces, thereby improving the overall visual quality of urban environments.