Attention approximation of mobile users towards their environment

Abstract

Public environments are increasingly equipped with interactive features, such as electronic maps for way finding, dynamic information displays or animated advertisements. Understanding the attention patterns of users within these environments is important for the design and evaluation of such interactive elements. We present a modelling approach based on dynamic adaptation of the users' field-of-view, bottom-up visual saliency calculation and task-dependent semantic interest modeling that allows approximating the attention of users with regard to a photorealistic 3D-model. This improved availability of attention information can help to design more usable navigation systems, identify problems for user groups with special needs and support the design of seamless attention switches between information elements.