Answer the question
In order to leave comments, you need to log in
How to identify speech?
Hello! I have the following task: There are audio tracks lasting from 30 minutes to 4 hours. You need to go through them and identify the speech of a particular person. My guess is that you need something like a neural network, which should be trained with patterns of cuts of this voice. Are there any libraries to implement this, and if I'm completely wrong about how to do this, please correct me. Preferably in js
Answer the question
In order to leave comments, you need to log in
try brain.js, if you need a js
neuron, it builds the most primitive, but it is easy to use
Well, yes, be prepared that it will all work slowly
PS on asm.js neurons "think" many times faster (v8 5.0.71, node 6.3.1)
https://www.google.ru/webhp?sourceid=chrome-instan... - google a convenient and licensed library or external API,
https://nodejs.org/api/addons.html - create a wrapper for libraries in the form of a native node module,
https://www.google.ru/webhp?sourceid=chrome-instan... - or look for ready-made wrappers.
Didn't find what you were looking for?
Ask your questionAsk a Question
731 491 924 answers to any question