Engineers from the University of Washington are working to optimize network bandwidth consumption during video calls. The aim is to bring the data rate down to 30kbps while optimizing video signals such that they are clear enough for deaf or hard-of-hearing users to decipher sign language.
Led by the MobileASL team, the video compression project increases image quality around the face and hands to better allow users to understand the semantics of the American Sign Language (ASL). The technology also uses the motion sensor to detect if the person is signing in order to reduce battery drain during video calls.
According to the team, they are developing ASL encoders that are compatible with the current H.264 standard using x264. These take into account visual and perceptual nuances that happen during ASL conversations. Although the project was developed to address the low bandwidth of the U.S. wireless networks, it could well be deployed by any cellular carrier.