I am home based medical transcriptionist, and was checking if it is possible to use Sphinx to aid my work. I have a set of audio files from a doctor, and the transcribed text. Can I use this for training? I tried reading through the documentation for SphinxTrain etc, but not being from a computer background, I am totally at a loss with all the technical terms etc.
Could someone please point me to a tutorial/document that describes how I can do this? That is, given a set of audio files from a person and the corresponding transcribed text, train the software to recognize that person's speech in subsequent files?
Thanks,
Anita.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Sorry, we don't offer an end-user medical transcription software. So you have two choices. Either you need to become a software developer and implement everything. Tutorial will help you get started then:
which mainly requires you to do simple manual work like segmentation and checking of the audio and checking the phonetic dictionary. You need to prepare a database - a collection of medical audio split on 3-5 sec chunks with appropriate transcription. Then you need to make this public/accessible to someone who could build help to build you a model.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
hi,
I am home based medical transcriptionist, and was checking if it is possible to use Sphinx to aid my work. I have a set of audio files from a doctor, and the transcribed text. Can I use this for training? I tried reading through the documentation for SphinxTrain etc, but not being from a computer background, I am totally at a loss with all the technical terms etc.
Could someone please point me to a tutorial/document that describes how I can do this? That is, given a set of audio files from a person and the corresponding transcribed text, train the software to recognize that person's speech in subsequent files?
Thanks,
Anita.
Sphinx is very difficult to use if you cannot program. I suggest you consider Dragon Naturally Speaking the medical transcription version.
Of course, still check out voxforge and submit your audio for the greater good of society :)
Hello Anita.
Sorry, we don't offer an end-user medical transcription software. So you have two choices. Either you need to become a software developer and implement everything. Tutorial will help you get started then:
http://cmusphinx.org/tutorial.html
Or, alternatively you could contribute to the VoxForge project:
http://voxforge.org
which mainly requires you to do simple manual work like segmentation and checking of the audio and checking the phonetic dictionary. You need to prepare a database - a collection of medical audio split on 3-5 sec chunks with appropriate transcription. Then you need to make this public/accessible to someone who could build help to build you a model.