12 - 15
The Fraunhofer Heinrich Hertz Institute HHI is a world leader in the development for mobile and optical communication networks and systems as well as processing and coding of video signals. Together with international partners from research and industry, Fraunhofer HHI works in the whole spectrum of digital infrastructure – from fundamental research to the development of prototypes and solutions.
At CeBIT 2018 Fraunhofer HHI presents the latest developments in Vision & Imaging Technologies and Video Coding & Analytics at Fraunhofer Booth E78, Hall 27, from June 12-15 in Hannover, Germany.
High-resolution 3D object and body caption for Mixed Reality interactions
The exhibit demonstrates Fraunhofer HHI’s latest developments in high-resolution 3D caption in the context of Mixed Reality human-technology interactions. The "Virtualizer" is able to capture objects in real time and in three dimensions with high resolution to display them directly in virtual scenes, which can be visualized by means of VR-glasses. With the Fraunhofer HHI algorithms for 3D body caption, tracking and gesture interpretation, VR users can interact naturally in the VR scene without disruptive controllers or markers and thus be represented as movable full-body avatars within the respective scene. The combination of the two technologies provides a unique solution for novel Mixed Reality interaction and collaboration scenarios such as remote assistance.
The Virtual Reality demo at CeBIT 2018 shows how a real object can be "handed" and explained to a user via the "Virtualizer" live in the VR environment.
Making Artificial Intelligence comprehensible
With the availability of large databases and recent improvements in Deep Learning methodology, the performance of AI systems reaches or exceeds the human level for an increasing number of complex tasks. Due to their nested non-linear structure, however, these very successful Machine Learning models are usually used in a black-box manner. This does not provide any information about what leads them to their predictions. Since this lack of transparency can be a major disadvantage in medical applications, for example, research into explainable AI (XAI) has recently become increasingly important.
Together with colleagues of the TU Berlin, scientists at the Fraunhofer HHI have developed a tool that allows such a tracking of even the most complex learning machines. This is the first step towards a transparent Artificial Intelligence.
Meet our experts at the booth of the Federal Ministry of Education and Research (Booth E52, Hall 27).