Menu

Pcoketsphinx decoder must tell when out of vocabulary word is spoken

Help
2018-08-28
2018-10-25
  • Farhan Ahmad

    Farhan Ahmad - 2018-08-28

    I have trained my acoustic model using sphinxtrain for some words that are written below

    YES NO ONE TWO THREE FOUR FIVE SIX SEVEN EIGHT NINE

    The accuracy of my model is quite good. Now I am using pocketsphinx to decode whatever I say or from audio file. The problem is when I say some word other than my vocabulary like my name farhan or wallstreet or something. the model decode this utterance to any one of the vocabulary word. But I want that the decoder in this case must say the spoken word is " out of vocabulary wor ". How can I do this?

    i have tried pocketsphinx keyword spotting. Made a keyword-list file and wrote these vocabulary words in it but not getting required results. i don't know what thresolds should be given for these single word phrases in the keyword-list.

    Help me!

     

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.