I was going to implement a speaker verification app for Android, and was wondering if there would be a way to get the voice features (MEL Coefficients) from the Android's Speech Recognition module?
Please note that speaker verification is slightly different than speech recognition, so that's why I'm not using the speech recognition directly.