I want to get the hypothesis only if the spoken word is “probably” within my vocabulary. Is there any way to neglect low confidence hypotheses or map them to a predefined value?

In other words, I do not want Pocketsphinx to map OOV words to wrong hypotheses.

Note: Only a single word may be OOV within a sentence. I do not want to ignore the whole sentence in that case.

Btw, I do not have any progress about my last question due to lots of other works I am working on, I will share here when I have any progress.

Thank you,

I’d have to recommend that you bring your Pocketsphinx functionality questions to the Pocketsphinx support board since they are the originators and maintainers of the code you have questions about. You’re always welcome to ask me questions about accessing Pocketsphinx functions you are already making use of in C or on the command line via Objective-C, but unfortunately I can’t offer general “how do I handle research task x with Pocketsphinx” support here since it is such a broad subject and one with its own support system.

Thanks,

Halle