0

I want to develop some web-based interactive activities for kids, and I want to use speech recognition as one kind of input. My question is not on how to record a sound in a browser, many other question already adressed that.

What I would like to know is whether there exists some kind of free online service that I could access through a dedicated API to do the speech recognition bit? My idea would be to record the sound and send it to a server, so that I don't have to set-up my own dedicated server just for the speech recognition part. This is how Siri works, for example. I assume Apple does not provide a free access to their speech recognition engine (at least not that I found), but maybe some other company does?

seb
  • 2,136
  • 3
  • 20
  • 27
  • 1
    http://stackoverflow.com/questions/4141893/speech-recognition-api that has a couple options. – AnthonyM Feb 01 '13 at 08:58
  • Children speech is quite different from elder people speech, so it's unlikely that ASR API will give you a good performance for children. You need to train a special children recording database to accurately recognize kids. One possible option for that is CMUSphinx http://cmusphinx.sourceforge.net. – Nikolay Shmyrev Feb 01 '13 at 13:42

2 Answers2

0

Google has launced a new api called the web speech api. check it out https://www.google.com/intl/en/chrome/demos/speech.html

The Third
  • 785
  • 2
  • 10
  • 30
0

I haven't found much either. There's a paper by Google that talks about training LSTM network on huge dataset to build a vocabulary for kids and adults and there is Pullstring that claims to have speech recognition engine targetted to kids but when it comes to API, I didn't find any. I'd appreciate any further research conclusion on this.

Sunil
  • 541
  • 2
  • 8
  • 17