Unspoken Words Recognition: A Review

Abstract

In recent years, unspoken words recognition has received substantial attention from both the scientific research communities and the society of multimedia information access networks. Major advancements and wide range of applications in aids for the speech handicapped, speech pathology research, telecom privacy issues, cursor based text to speech, firefighters wearing pressurized suits with self contained breathing apparatus (SCBA), astronauts performing operations in pressurized gear, as a part of communication system operating in high background noise have propelled words recognition technology into the spotlight. Though early words recognition techniques used simple maximum likelihood algorithms only but the recognition process has now graduated into a science of mathematical representations and comparison processes. This survey paper provides an up-to-date review of the existing approaches and offers some insights into the study of unspoken words recognition. A number of typical techniques and EMG based approaches are discussed in this paper. Furthermore, a discussion outlining the incentives for using recognition techniques, the applications of this technology, and some of the difficulties plaguing the current systems with regard to this topic have also been provided. Keywords-Speech Pathologies, Electromyography, Hidden Markov Models, Myoelectric Signals (MES).

2 Figures and Tables

Cite this paper

@inproceedings{Saini2013UnspokenWR, title={Unspoken Words Recognition: A Review}, author={Manish Kumar Saini and Jagbir Singh Saini}, year={2013} }