Certain investigations on performance analysis of fobust automatic speech recongnition system using voice activity detection and speech enhancement algorthim

Abstract

newlineSpeech constitutes the primary form of human communication newlineSpeech has the potential to be a better interface other than keyboard and newlinepointing devices A speech interface would support many valuable newlineapplications like telephone directory assistance hands busy applications in newlinemedicine office dictation devices etc This has motivated research in newlineAutomatic Speech Recognition ASR newlineIn recent years the progress of voice signal s parameter extraction newlinetechnology promoted the rapid development of speech recognition Automatic newlineSpeech Recognition ASR in noisy environments remains a challenging newlineproblem since there are many possible types of environmental distortion and newlineit is difficult to compensate for all of these distortions accurately The primary newlinereason for poor recognition performance in noise is the mismatch between newlinetraining and test conditions newline newline newline newline