Speaker verification using Gaussian Mixture Model (GMM).

---

Run

1. Run in google colab

Plots

Results

accuracy score for DR3_FPKT0: 100.0%
eer score for DR3_FPKT0: 11.377245508993903% with threshold -7.957904145487239
              precision    recall  f1-score   support

           0       1.00      1.00      1.00       167
           1       1.00      1.00      1.00        10

    accuracy                           1.00       177
   macro avg       1.00      1.00      1.00       177
weighted avg       1.00      1.00      1.00       177


accuracy score for DR3_MJJG0: 98.87005649717514%
eer score for DR3_MJJG0: 13.7724550898456% with threshold -7.269569053999492
              precision    recall  f1-score   support

           0       0.99      0.99      0.99       167
           1       0.90      0.90      0.90        10

    accuracy                           0.99       177
   macro avg       0.95      0.95      0.95       177
weighted avg       0.99      0.99      0.99       177


accuracy score for DR3_FCMH0: 99.43502824858757%
eer score for DR3_FCMH0: 1.1976047904198035% with threshold -6.315210336766802
              precision    recall  f1-score   support

           0       0.99      1.00      1.00       167
           1       1.00      0.90      0.95        10

    accuracy                           0.99       177
   macro avg       1.00      0.95      0.97       177
weighted avg       0.99      0.99      0.99       177


accuracy score for DR3_MWJG0: 99.43502824858757%
eer score for DR3_MWJG0: 10.000000000018641% with threshold -7.284514182552126
              precision    recall  f1-score   support

           0       0.99      1.00      1.00       167
           1       1.00      0.90      0.95        10

    accuracy                           0.99       177
   macro avg       1.00      0.95      0.97       177
weighted avg       0.99      0.99      0.99       177


accuracy score for DR3_MTAA0: 99.43502824858757%
eer score for DR3_MTAA0: 4.790419161838868% with threshold -6.471338153605574
              precision    recall  f1-score   support

           0       0.99      1.00      1.00       167
           1       1.00      0.90      0.95        10

    accuracy                           0.99       177
   macro avg       1.00      0.95      0.97       177
weighted avg       0.99      0.99      0.99       177


------------------------------------------------------------------------------------
------------------------------------------------------------------------------------
------------------------------------------------------------------------------------
Final average test accuracy: 2.959375840731773%
Final average test EER: 0.244867408042362%
Final accuracy score for all speakers: 99.43502824858757%
Final eer score for all speakers: 9.131652661064422%
              precision    recall  f1-score   support

           0       1.00      1.00      1.00       835
           1       0.98      0.92      0.95        50

    accuracy                           0.99       885
   macro avg       0.99      0.96      0.97       885
weighted avg       0.99      0.99      0.99       885

...

Support

Tested with: python3.6 python3.7 python3.8

TO-DO:

• gmm-ubm
• svm
• inference script