A research team led by scientists at the University of Buffalo modified a headset with a noise-cancelling system that allowed the electronic device to "see" and translate sign language when paired with a smartphone.
The system, dubbed SonicASL, performed 93.8% in tests conducted indoors and outdoors and included 42 words. Examples of the vocabulary used were 'love', 'space' and 'camera'. Under the same conditions, using 30 simple sentences - e.g. "Nice to meet you" - SonicASL showed a 90.6% efficiency.
More work needs to be done before the technology can enter the market as a commercial product, the researchers stressed. For example, the vocabulary of SonicASL needs to expand significantly. In addition, the system must be able to read facial expressions, which is an important component of sign language.
The researchers intend to continue working on the system. Their main motive is that there are more than 72 million people worldwide with hearing impairments. All of them have difficulty communicating with the rest of the people. The task is very challenging because deaf people around the world deal with more than 300 different sign languages.
According to the researchers, the invention uses Doppler technology to "sense" tiny vibrations or echoes in the acoustic waves created by the movement of the hand. The research will be presented at the ACM Conference on Ubiquitous Computing (UbiComp), September 21-26.
Source: acm.org, buffalo.edu