Automated support system designing for people with limited communication
Artem Goncharenko, Lilia I. Voronova, Vyacheslav I. Voronov, Aleksey A. Ezhov, D. V. Goryachev · 2018
The article describes designing, modeling and testing of automated gesture recognition system, with access and control from wearable mobile devices. It will help people with disabilities to communicate with people who do not understand gesture language. For example, during a painful attack on Board the aircraft to ask doctor help. The developed system consists of subsystems that perform tasks of authorization, recording, storage and conversion of video, sending data to the server and recognition. The neural network with error back propagation is used for recognition. The topology of the used neural network is the perceptron. This is a prototype system that works under light conditions, and in the future will be refined.