A Neuro-Linguistic Model for Speech Recognition in Tone Language

Abstract

The primary aim for this work is to develop a speech recognition system that exploits the computational paradigm with learning ability and the inherent robustness and parallelism in ANN coupled with the capability of fuzzy logic to model vagueness, handling uncertainness and support for human reasoning. This research made two unique contributions to the area of speech recognition. Firstly, we identified and specify entities in the speech recognition process using a neuro-linguistic approach. This approach not only offers a non-conflict degree for different speech recognition situation, it also provide stable integration mechanism for various level of speech unit. Secondly, this work is the first to illustrate speech recognition model using an African language, in this case Yoruba language. It is important, however, to note that this approach can also be used for any tone language and with little modification for all languages as well.


(Botswana Journal of Technology: 2003 12(2): 32-39)