| 摘要 |
<p>A speech synthesizer (3) produces prompts in the voice of a native speaker of the language to be learned, which the student may imitate or reply to, and a phrase recogniser (1) which uses keyword recongnition is employed so that the system understands spoken phrases and interactive dialogue may take place. The student's progress is monitored by measuring the deviation from his original speech recognition template; when this difference is sufficiently large that the recogniser (1) can no longer recognise what the student is saying, the system re-trains and updates the template. In another embodiment, the system includes a display which shows the native speaker's mouth shape whilst the words to be imitated are spoken by the speech synthesizer (3); and a video pick-up and analyser for analysing the shapes of the student's mouth to give the student visual feedback.</p> |
