DEVELOPING A SYSTEM FOR ASSESSING THE CORRECT PRONUNCIATION OF WORDS IN UZBEK LANGUAGE

Abstract

This work presents the stages and results of implementing a system for assessing the correct pronunciation of words in Uzbek language. Filtering, segmentation, spectral representation, feature extraction, and evaluation based on the DTW algorithm were performed on the recorded speech signal. During the testing of the results, the pronunciation of the same word was 77.50%, and word recognition was 100%. Based on the results, users had the opportunity to work on themselves and improve their pronunciation accuracy