The recognition of gestures requires in contrast to face recognition analysis in motion. To make this reliable we collaborate with our partners form university Düsseldorf in various research projects.
Within the projects we put our effort in the recognition of selected elements of the German Gesture Recognition Language (DGS) supported by neutral networks. In addition we work on a common modern sensor platform to allow the recognize the depth of objects.
The video shows a technology preview from our current work area "gesture recognition". In contrast to image recognition, a movement has to be largely completed before it can be recognized. Therefore there is always a time delay in gesture recognition. In addition to the hands, the face plays a central role in the analysis of the DGS.