Automatic Speech Recognition in the Assessment of Child Speech
Loridana Buttigieg, Helen Grech, Simon G. Fabri, James Attard, Philip Farrugia · 2021
In fact, researchers have been investigating ways of automating the process of phonetic transcriptions, for example by utilizing speech recognition algorithms and automatic speech recognition (ASR) systems. ASR systems require algorithms to analyse pauses between syllables, relative syllable stress between strong and weak syllables and phonemic accuracy devices which recognize the child&s;s speech as correct or incorrect. The accuracy levels of the ASR system during the initial assessment were compared to those of the second assessment in order to determine any improvements. The findings of this study also conform to the findings of Dudy, Bedrick, Asgari, and Kain who found that the accuracy of the ASR system improved significantly through specific models aimed to improve system performance. Hence, more efforts to improve the Maltese-English ASR system are required and more recordings need to be collected to train the ASR system more extensively.