Recognizing multiple peoples voices

问题

I am looking for an open source voice recognition engine that, instead of responding to spoken words, can determine who is speaking. Does anyone know where I might be able to find something like this?

回答1:

You can consider

Bob SPEAR

https://pypi.python.org/pypi/bob.bio.spear

Alize/Mistral

http://mistral.univ-avignon.fr/index_en.html

GMM speaker identification in matlab

https://github.com/codyaray/speaker-recognition

Very basic speaker recognition in Java, not really accurate

https://github.com/amaurycrickx/recognito

In python, quite a state of the art toolkit

https://pypi.python.org/pypi/SIDEKIT

回答2:

What you are looking for is speaker diarization and identification.

You can use Microsoft Speaker Recognition for doing the speaker identification task: https://www.microsoft.com/cognitive-services/en-us/speaker-recognition-api#identification

回答3:

If you're Symfony2 developer, I provide you bundle for Microsoft Speaker Recognition API: https://github.com/onekit/speaker-recognition-bundle

Feel different between speaker and speech. Speech-To-Text can be recognized by Chrome Browser.

来源：https://stackoverflow.com/questions/32914148/recognizing-multiple-peoples-voices