DefinePK

Abstract

Being speech the most natural way of communicating among humans, it should be possible to use it without complications in any aspect of a human live. Unfortunately this is not possible in certain situations like in inappropriate environments or speaking disable people. Unvoiced speech recognition is capable of solving these issues through the acquisition of biological signals directly related to speech. Consequently, an analysis of silent speech recognition systems using electromyographic signals is presented. Applications of this technology in medicine, human interfaces, voiced and unvoiced recognition are showed. A description of hardware and software used in EMG-based projects is realized, along with an introduction to multiple techniques used for feature extraction and classification of myograhpic signals. The results obtained by the different projects are analyzed and the main difficulties still present in this kind of systems are commented.