Оценка качества произношения на основе метода нечеткого фонетического кодирования
The paper is devoted to the automatic assessment of phoneme pronunciation quality for computer assisted language learning systems. The novel pronunciation training algorithm is proposed. In this algorithm, at first, a student has to achieve a stable pronunciation of all sounds by using the phonetic database from ideal speaker. The second, novel, stage of our algorithm is the phoneme recognition with the modification of fuzzy phonetic coding method. The experimental results are provided for learning of isolated sounds in Russian and English languages.