| 摘要 |
PURPOSE: A pronunciation evaluation device according to a level of pronunciation and a method thereof are provided to inspire a learner by providing a pronunciation evaluation result by variably updating a threshold value according to the level of pronunciation. CONSTITUTION: A voice feature extraction unit(110) extracts voice feature information from a voice signal recognized in a voice recognition unit(100). A voice reliability measurement unit(120) measures a current reliability value of the voice signal using the extracted voice feature information. A threshold update unit(130) updates a threshold value based on a compared result between the measured current reliability value and previously measured reliability value. A pronunciation evaluation providing unit(140) compares the updated threshold value and the current measured reliability value and provides a pronunciation evaluation result. [Reference numerals] (100) Voice recognition unit; (110) Voice feature extraction unit; (120) Voice reliability measurement unit; (130) Threshold update unit; (140) Pronunciation evaluation providing unit; (AA) Sound signal input |
