Based on a new type of award-winning sensor technology, the time-of-flight (TOF) camera, ARTTS will develop algorithms and prototype systems to solve open computer-vision problems and enable new applications that involve multimodal interfaces and the sensing of people and their actions. Unlike a conventional video camera, the TOF camera delivers not only an intensity image but also a range map that contains a distance measurement at each pixel, obtained by measuring the time required by light to reach the object and return to the camera (time-of-flight principle).

Prototype ARTTS project camera

Intelligent recognition and tracking solutions developed for TOF cameras will provide a crucial enabling technology for multimodal interfaces. The current state-of-the-art 3D TOF systems will be considerably improved so as to increase the depth resolution and quality of the signal and to reduce size, power consumption, and cost. A prototype of the improved system will be developed and released within the ARTTS project.

Advanced signal processing methods will increase the quality of the signal and derive the features needed for the subsequent development of algorithms that can track people and recognize their activity. Software toolboxes for object tracking and action recognition will be developed that can serve as a basis for various applications.