in my application
what is recognized with SpeechRecognizer it is written in the label.
I wish that if the label to recognize couples words even if they have a word like, for example:
1) home-sky
2) home-land
3) home-sky-earth
as I can do?
what is recognized with SpeechRecognizer it is written in the label.
I wish that if the label to recognize couples words even if they have a word like, for example:
1) home-sky
2) home-land
3) home-sky-earth
as I can do?
--
You might not be able to use the Speech Recognizer to do exactly this with any compiler. What the speech recognizer sometimes 'understands' is not what is said. The Google SpeechRecognizer works with is very imperfect.
What have you done so far? Have you already coded to get your 'label' to recognize simple works like "cat" or "dog".
Here are some blocks (untried) that might help:
A construction like this might allow you to 'filter' the results of what the speech recognizer "understands.".
Does it work for you? Another possibility is for the user (speaker) to enunciate homeland with a slight pause so what he/she actually says is home land ? It is difficult to get the speech recognizer to do complicated things. Perhaps this technique will work for the app you have in mind.
--
댓글 없음:
댓글 쓰기