Low-cost indoor localization using cameras

Many systems have been proposed in recent years that provide for the tracking and localization of users in indoor environments, often with a specific focus on pervasive computing settings. Our solution AmbiTrack, as presented here, allows for a marker-free localization and tracking of multiple persons, meaning that users are not required to carry special items or tags with them in order for the system to work. This approach allows for an application of AmbiTrack in circumstances where wearing a tag is not viable, e.g., in typical Ambient Assisted Living scenarios where the users of the provided technological systems are usually not technologically well-versed. In this contribution, we explain AmbiTrack and also introduce the adaptations we made for the 3rd EvAAL competition of 2013 in order to make the system more reliable in tracking multiple persons, using context information for improving the recognition rate, and for simplifying the set up and configuration process.