Korean speech signals can be divided into the phonemes such as consonant, vowel, and end-consonant according to the characteristic and structure its pronunciation.
The experiment is carried out by dividing 975 isolated words into phonemes. The
rec...
Korean speech signals can be divided into the phonemes such as consonant, vowel, and end-consonant according to the characteristic and structure its pronunciation.
The experiment is carried out by dividing 975 isolated words into phonemes. The
recognition rate of consonant, vowel, and end-consonant with DTW are 85.7(%), 80.0(%) and 93.1(%), respectively. The recognition rate of consonant, vowel, and end-consonant with HMM are 82.1(%), 77.7(%) and 86.9(%), respectively.
It is shown the Korean speech signals are recognized in the same way minimum data and short processing time.