Study on Visual Perception Patterns of Public Space in Urban High-Density Residential Areas: The Example of Xi’an

Abstract

In response to the accelerated process of urbanisation, the development of high-density residential areas has emerged as an effective strategy to meet the increasing demand for housing. However, the limited public space available in these areas frequently fails to provide adequate separation between public and private domains. The proximity between public and private spaces often results in residents' sightlines intersecting with private areas during public activities. Such spatial arrangements compromise both personal privacy and the quality of public activities. This study examines the relationship between the proportions of spatial elements within the human field of view and the duration of gaze across a range of settings. The objective is to analyse visual attention patterns to inform strategies to enhance the attractiveness and engagement of public spaces. In contrast to conventional design approaches that rely on subjective evaluations, the measurement of human physiological data provides a more objective means of quantifying spatial experiences. Given the crucial role of visual information in spatial perception, eye-tracking technology captures and quantifies attention patterns, enabling designers to provide precise and actionable recommendations for improving spatial arrangements. This study employed a 360-degree panoramic camera to capture 120 photographs of public activity areas across 20 residential communities in Xi'an, Shaanxi. Semantic segmentation was used to extract information from the images. Furthermore, 50 participants observed the scenes with their eye-tracking data recorded. By correlating spatial data with eye-tracking metrics, this study investigates the relationship between spatial elements and visual attention in residential public areas, revealing perception patterns across different spaces.