Understanding human-place interaction from tracking and identification of many users

This paper considers the problem of understanding human-place interaction, such as relationships among many users in a space and interactions between users and their surroundings, from trajectories of users in a common space. The discovered information can be applied to provide a number of services. For example, we can determine the optimal arrangement of items in a store or at an exhibition to maximize the profit or attention and systematically manage the pedestrian traffic. Users in a space is detected and tracked by a vision-based multi-target tracking algorithm and trajectories of users are identified by combining visual information and accelerometer readings from users' smartphones. We demonstrate that trajectories of users can be used to reveal a number of useful information about the users and the space, such as spatial occupancy of individual users, intimacy between users, objects of interests, and a common interest of users.