Desarrollo de interfaces para la detección del habla sub-vocal

Abstract

This paper explores the most important techniques currently used to detect sub-vocal speech in people with cerebral palsy as well as for commercial purposes, (e.g. allow communication in very noisy places). The methodologies presented deal with speech-signal acquisition and processing. Signal detection and analysis methods are described throughout the whole speech process, from signal generation (as neural impulses in the brain) to the production sound in the vocal apparatus (located in the throat). Acquisition and processing quality depends on several factors that will be presented in various sections. A brief explanation to the whole voice generation process is provided in the first part of the article. Subsequently, sub-speech signal acquisition and analysis techniques are presented. Finally, a section about the advantages and disadvantages of the various techniques is presented in order to illustrate different implementations in a sub-vocal speech or silent speech detection device. The results from research indicate that Non-audible Murmur Microphone (NAM) is one of the choices that offer huge benefits, not only for signal acquisition and processing, but also for future Spanish language phoneme discrimination.

Keywords

• speech sub-vocal
• interface
• speech recognition