I need to make a fast and accurate recognition of the alphabet (letters) using
Sphinx 3.
I never worked with it so I'd be thankful for both good tutorial on sphinx 3
and an example of letters recognition.
Thanks.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
try to follow [http://www.speech.cs.cmu.edu/15-492/assignments/asr/index.html]
(http://www.speech.cs.cmu.edu/15-492/assignments/asr/index.html) it's not
alphabet recog but you will get general idea about how to use sphinx
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
argument to decoder. But the same thing (phoneme recognition) can also be done
with other modes (fwdflat or fwdtree (which is the default mode)). Would there
be a difference between these modes (as in would allphone give better results
as compared to fwdtree/fwdflat)? I used to use default mode with dictionary
just like phone.dict given in that tutorial.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Is there some example project for Sphinx 3, something like "Hello world"
sphinx3 application?
There's one for Sphinx4 but i can't find anything like that for Sphinx3.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hello,
I need to make a fast and accurate recognition of the alphabet (letters) using
Sphinx 3.
I never worked with it so I'd be thankful for both good tutorial on sphinx 3
and an example of letters recognition.
Thanks.
Are you talking abt handwriting recognition or speech recognition?
speech recognition
have you installed sphinxbase, sphinx3, sphinxTrain, cmucltk etc?
try to follow [http://www.speech.cs.cmu.edu/15-492/assignments/asr/index.html]
(http://www.speech.cs.cmu.edu/15-492/assignments/asr/index.html) it's not
alphabet recog but you will get general idea about how to use sphinx
There is a good enough tutorial in the project's wiki: http://cmusphinx.sourc
eforge.net/wiki/phonemerecognition
@vpanayotov In that tut they use
argument to decoder. But the same thing (phoneme recognition) can also be done
with other modes (fwdflat or fwdtree (which is the default mode)). Would there
be a difference between these modes (as in would allphone give better results
as compared to fwdtree/fwdflat)? I used to use default mode with dictionary
just like phone.dict given in that tutorial.
See the following thread: http://sourceforge.net/projects/cmusphinx/forums/fo
rum/5471/topic/1927422
Thanks for the link vpanayotov!
Thank you all. I really appreciate your help. I'll check those links.
Is there some example project for Sphinx 3, something like "Hello world"
sphinx3 application?
There's one for Sphinx4 but i can't find anything like that for Sphinx3.