Wayfinding Design Toolkit with Real-time Prediction of Visual Saliency as an Iterative Feedback

Abstract

Iterative design is particularly important in the urban planning process, as they are inherently irreversible and costs a lot. However, novice designers find difficulty in the iterative design process as they are unsure of the quality of their designs. They have trouble in the iterative design process and expecting where people will pay their attention, which is crucial in designing wayfinding signs and landmarks for city navigation. In this paper, we introduce Urban Salviz, a tool that visualizes real-time saliency prediction of city environment from pedestrian’s view to support iterative design process of novice designers. We applied our tool to a game engine for spatial interaction, but it can be integrated with any computer-aided design program with proper configuration of communication. Our user study result indicates that the tool brings more helpful and efficient design experience without sacrificing overall usability. The result of crowd-powered design evaluation shows that the utilization of Urban Salviz helps people produce designs that attract eyes without sacrificing aesthetic values and integrity in surrounding environments. We believe our tool can lower the barrier for urban planning, allowing more group of designers to participate, especially for landmarks and easily recognizable street layout. The tool can potentially expand user groups to enable the general public to potentially shape their daily interactions in the city and thereby, can eventually contribute planning of more inclusive cities with eye-attracting wayfinding signs that reduce confusion and helps city navigation.