Small logo of ETH main building ETH Zurich : Computer Science : Pervasive Computing : Distributed Systems : Education : Student Projects : Abstract

Device Recognition for WoT Interaction (B)

Status: Abgeschlossen

In our efforts to create an ecosystem of smart things that communicate and interact, we are currently deploying prototypes of devices (e.g., smart ther- mostats, electricity meters, environmental sensors). We expect that, in densely populated environments of such devices, an end user is prone to lose track over which devices communicate with each other and where data produced by a specific device is sent and analyzed.

The goal of this project is to use visual object recognition technologies for ef- ficient direct interaction with smart things. Following up on a previous bachelor thesis that was carried out during FS2013, this project shall in particular tackle the issues of improving the efficiency and usability of the handheld interface. This could, for instance, be done by taking into account context information (e.g., the user location). Additionally, we want to improve the selection and tracking of multiple objects within the camera frame. The idea here is that robust multiple-object tracking would enable a wealth of intriguing applications for smart environments, such as the real-time visualization of data exchanges between objects.

Tools / Languages: Java, Android, OpenCV, C++

Student/Bearbeitet von: Sezer G├╝ler
Contact/Ansprechpartner: Simon Mayer

ETH ZurichDistributed Systems Group
Last updated October 20 2014 03:42:39 PM MET sm