Skip to content

How to use GaussianMixture.java to create speaker model #2

Description

@satwickdash

Hi @wahibhaq

Can you please explain how can I use the GaussianMixture.java class, in your android project of speaker detection ?

I tried to import your project into Android Studio, and now the extraction of MFCC features and dumping them into a CSV file, works perfectly. But I cannot understand how to proceed to create a Gaussian Mixture Model from those features.

I tried looking up on what should be provided as weights to the GaussianMixture class, and found the following:
StackExchange - Covariance and Weights
StackOverflow - Gaussian Component Class' weight

From what I understand, is that K-Means clustering would work to cluster different speakers separately, when the MFCC features of audio samples of all the speakers are available.
But how will K-Means clustering work, when audio sample is collected for a particular speaker ?

Please guide if I'm proceeding in the right direction, and correct my understanding if needed.
Thanks.

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions