| 摘要 |
The isolated speech word is extracted from an input speech signal. First, the speech signal is divided into a high frequency band signal and a low frequency band signal, the power levels of each of the high and low frequency band signals are independently compared with respective threshold levels, and the front and the end of the speech word are detected through these independent comparisons. An apparatus for extracting the isolated speech word comprises: a first means (11) for receiving an input speech signal and detecting an average consonant power therefrom; a second means (12) for receiving said input speech signal and detecting an average vowel power therefrom; and a third means (13) for extracting an isolated speech word from said input speech signal by using the thus detected average consonant power and the thus detected average vowel power. |
