vSign – photo-realistic sign language video generation for broadcast sign language interpretation

Presenter(s): Richard Bowden (University of Surrey), Robin Ribback (SWISS TXT), Necati Cihan Camgoz (University of Surrey), Ben Saunders (University of Surrey)

Avatar technology has been routinely rejected by the Deaf community, but imagine if AI could be used to provide more human interpretation for the Deaf and increase the ease of work and throughput of the interpreter. vSign is a photo-realistic sign language video generation system that is able to generate high-quality sign language videos. It does not replace the interpreter, but frees the interpreter from the burden of the studio and also maintains their anonymity.