Abstract
In this paper, we present a method for automatic detection of word pronunciation errors. This method is based on Automatic Speech Recognition (ASR) technologies and uses a combination of two acoustic modeling methods: Hidden Markov Models (HMM) and Vector Quantization (VQ). The probability-based normalization technique evaluates the average quality of word pronunciation and shows any deviation in time. This method produces acceptable results for 10 Russian digits. The conformity between experts' and system's estimations is approximately 73%.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have