A glove that translates American Sign Language in real time
Translating sign language into speech in real time could soon be a reality, thanks to a pair of smart gloves created by bioengineers at the University of California, Los Angeles (UCLA), USA. Made from normal gloves with slim sensors of electrically conducting yarn that stretch along the fingers, the sensors detect hand motions and finger placements that each stand for letters, numbers, words and/or phrases.
The signals are then transmitted to a circular-shaped circuit board located at the wearer’s wrist, before these again are transmitted, wirelessly, to a smartphone that translates the movements. The processing speed is said to be about one word per second.
The materials used include stretchable polymers that are both lightweight and low-cost, whereas alternative wearable devices can be bulky, heavy and uncomfortable, shares Jun Chen, Assistant Professor
of Bioengineering at UCLA.
‘The core of the sensing unit is composed of a conductive yarn coiled around a rubber microfibre, with the entire body sheathed by a polydimethylsiloxane sleeve,’ Chen explains.
During tests, four adhesive sensors were affixed to participants’ faces, in-between the eyebrows and on one side of the mouth, to capture facial expressions. The testers repeated every movement 15 times, before a machine learning algorithm, especially created for this project, adapted these gestures into their respective letters, numbers or words. The system currently recognises 660 signs.
A patent has been filed, but Chen explains that a commercial model will need to translate quicker with added vocabulary.