A visual salience model for wayfinding in 3D virtual urban environments

- Two variables in wayfinding, global salience and local salience, are proposed.
- The weight of each variable is determined based on difference evaluation method.
- Salience measuring model of buildings is designed.
- 3D building models with different details are created according to salience value.

The study develops a modeling framework to quantify the saliency of individual buildings and to visualize them accordingly in a virtual 3D urban environment for wayfinding. Based on the cognitive characteristics of wayfinding behaviors and considering the relationships among pathfinder, buildings, and local context of the urban environment, this study defines and distinguishes two types of variables related to the saliency of a building, the global salience variables and the local salience variables. Then an adaptive weight distribution method is developed to determine the weight for each variable. Global salience is estimated in the context of a broader urban space, while local salience is estimated in the local region. Finally, a synthesized salience model is constructed by integrating the global and local salience values with consideration of the visibility factor. The salience values are used to identify most salient landmarks in each scene on the path of wayfinding. In the implemented prototype, 3D visualization is made for each scene showing buildings with varying levels of details according to their salience values. This study contributes to the research on intelligent 3D navigation systems by providing effective visual guidance for wayfinding tasks in virtual urban environments.