International Journal of Soft Computing
ISSN: OnlineISSN: Print 1816-9503
Abstract
The accuracy of input speech signal is very important as speech signal is now considered as an input source for several business domains.. This study attempts to propose a usability standard procedure for speech recognition. It also proposed an algorithm to investigate the cause for errors in recognition by including several related research works in this area and suggest an error recognition and repair procedures. The research suggests that it is practically possible to predict the misrecognized utterances with a high degree of accuracy from an utterance’s sound file, the language model being employed and recognizer outputs such as confidence, In addition, there is empirical data upon which to base successful repair strategies in relation to these misrecognitions.
How to cite this article:
E. Ramaraj and E. Chandra . Speech Recognition Standard Procedures, Error Recognition and Repair Strategies.
DOI: https://doi.org/10.36478/ijscomp.2007.88.95
URL: https://www.makhillpublications.co/view-article/1816-9503/ijscomp.2007.88.95