The use of machine learning and artificialintelligence will help in the near future to integrate people with various physical disabilities into society. The developers of the University of California, Los Angeles (UCLA) have created a system that allows you to remove the communication barrier between people who are forced to use sign language and other people. The device in the form of an electronic contactless glove is capable of translating the translation of the American sign language into English, almost in real time.
The system consists of a pair of gloves with thin,flexible sensors that can control the work along the entire length of each of the five fingers. These sensors, made of electrically conductive filaments, record the movements of the hands and the positions of the fingers, which indicate individual letters, numbers, words and phrases.
The device converts finger movements toelectrical signals that are sent to the control board the size of a dollar coin worn on the wrist. Further, these signals are transmitted wirelessly to a smartphone, which translates them into words at a speed of about one word per second. Then the application speaks the translated phrase through the speaker.
Overhead sensors are added to the system, mounted on the user's face on both sides of the mouth and between the eyebrows. This is necessary to capture facial expressions, which is part of the American sign language system.
Autonomous work of a glove is providedthe battery. In the process of preparation, after working using machine learning and the experience of four deaf Americans, the algorithm began to confidently recognize 660 characters, including letters and numbers. In the future, according to engineers, the algorithm will actively replenish its vocabulary and increase the translation speed. Currently, the invention is in the process of registering a patent.