CODEWITHSUNDEEP
androidspeech-recognitioncmusphinxpocketsphinxphoneme
user5794813
user5794813

Reputation:

Poor recognition accuracy of Pocketsphinx using phoneme recognition on Android, French language

I am working on a project where I have to integrate the speech functionalities of Pocketsphinx into an android application. In fact, I have to integrate the phoneme recognition functionality provided by Pocketpshinx that should be able to recongize phonemes in French language, e.g. the speech recongnizer should be able to recognize syllabes( like "de", "re", "se", etc), consonants (like "m", "f", "g", etc), double-consonants(like "kl", "ks", "gr",etc) and vowels(like "a", "o","e",etc). Right now, I have integrated the Pocketsphinx for recognizing the phonemes mentioned above, but I have really bad results. For example, when I pronounce the "o", the recognized result sometimes is: "SIL ff ei au" (even I did not pronounce the letter "f" and "e" at all), or something else is appeared at the beginning that is not pronounced. The letters that are appeared at the beginning are not always the same (sometimes I get "ll", "uu", etc). But sometimes, the letter I pronounce appears at the beginning (e.g. for "a", I get "SIL aa SIL") , but this happens really rarely. So, could you guys please help and let me know what could be the problem and any suggestions for solving this problem? Do I need to create a language model and a dictionary only with these letters that I am interested in (since the number of letters is fixed)? Or, maybe I need to do the adaption of the acoustic model, or train the acoustic model?

Thank you very much in advance!

Upvotes: 1

Views: 590

Answers (1)

Ayush Jain
Ayush Jain

Reputation: 144

In your model find feat.params and change -cmuinit value according to your choice, Also, the value of "-kws_threshold" to 1e-40f for removing false alarms.

Upvotes: 1

Related Questions