I am working on speaker identification problem using GMM (Gaussian Mixture Model). I have to just identify one user present in the given audio, so for second class noise or silent audio may use or not just like in image classification for an object we create a non-object class.

I have used a silent class is always showing the user is present ( which is not).

If any other model can give better accuracy fulfil the condition that only 30 sec of audio of a particular user is available and given test audio may has long size.

  • $\begingroup$ Why did you choose GMMs? Any intuitions ? Why not Neural Nets? $\endgroup$ – Sharan Jan 10 '20 at 9:19
  • $\begingroup$ The speaker files will be maximum 30seconds and 3-5 files only (for training). Since GMM is a statistical model as a statistical model works better with small data compare to the neural network. $\endgroup$ – Posi2 Jan 10 '20 at 9:40

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.