Menu

What's my next step while developing a Speech Recognition system?

2017-07-12
2017-07-12
  • Mehmet Mert Yildiran

    Hi, I'm trying to develop (and at the same time learn how a) Speech Recognition system (works). I'm currently able to draw the spectrum and find the peaks.

    The current state of my program is in here (there is not much, just ~200 lines). Also I recommend you to read the README.md (it's not long, a few lines).

    My question is what's my next step? I think my next step is trying to recognize the phonemes. But I'm not sure how can I do that? How can I distinguish the vowels 'a' from 'e' for example. Also how can I detect the boundaries (or transitions) between the phonemes. From this point, do I need to jump into Hidden Markov model or Neural Networks? Also what data should I feed to HMM or ANN, the peaks? or the spectrum graph's itself?

     

    Last edit: Mehmet Mert Yildiran 2017-07-12
    • Nickolay V. Shmyrev

      My question is what's my next step?

      Read speech recognition textbook.

       
      • Mehmet Mert Yildiran

        I already did. I read a lot of academic material. But most of the time they have created more questions than answers, in my mind...

        When it comes to practice, people like me (alien to this research because of certain reasons) can stuck and need guidance. So I'm asking for a short guide, a To-Do List.

         

Log in to post a comment.

Want the latest updates on software, tech news, and AI?
Get latest updates about software, tech news, and AI from SourceForge directly in your inbox once a month.