A team of bioengineers at the University of California Los Angeles (UCLA) has developed an electronic non-contact glove that is capable of converting gestures into speech.
The glove is equipped with stretchable electrical sensors made of a special material that work along the entire length of the fingers, tracking their movements. The system is complemented by sensors that can be stuck on the face (between the eyebrows and on one side of the mouth) to also “read” facial expressions, which is part of American Sign Language.
The gadget is powered by a battery and a wireless network and can transmit information about movements to a smartphone, where a special application translates sign language at a speed of one word per second. The system currently recognizes about 650 characters, including all letters and numbers, thanks to machine learning technology. The speaker then pronounces the translated word of the gesture language clearly.
According to the researchers, the commercial version of the system will have an even larger vocabulary and significantly increase the rate of translation.