Navigation auf uzh.ch
Sign languages are characterized by the use of multiple production modalities, making them inherently multimodal: In addition to the more familiar manual features – the activity of the hands and arms –, there are non-manual features, i.e. static and dynamic features of the upper body, head and face (eyebrows, gaze, eyelids, nose, mouth). These features are equally significant from a linguistic point of view. Incorporating non-manual features into automatic sign language processing (sign language recognition, translation, and synthesis) forms one of the research challenges we are dealing with. In addition to the core processing tasks, our group also investigates computer-assisted sign language assessment.
Note that, contrary to what is sometimes reported in the media, the state of sign language technology is not such that, e.g., fully automatic sign language translation in an unrestricted domain would be possible.