-
Notifications
You must be signed in to change notification settings - Fork 5
Research
Chad edited this page Sep 23, 2017
·
11 revisions
-
Automatic Speaker Diaratization using Machine Learning Techniques - Trained 3 models (k-means, SVM, DNN), recommendation for SVM, deeper DNN and RNN.
-
STFT: Short Time Fourier Transform using Tensorflow - Windowing of audio sample, applying FFT to segments (Tensorflow implementation here).
-
End-to-End Speaker Verification - Use of i-vectors and d-vectors as inputs to DNN.