-
Notifications
You must be signed in to change notification settings - Fork 0
Speaker Recognition
arvind0422 edited this page Jul 5, 2017
·
4 revisions
The built system could classify 10 voices with 100% accuracy tested on the ELSDSR dataset with 90 seconds of test speech per speaker.
- Confusion Estimate.
- Create a threshold to identify if the speaker is new.
- User Interface to Train & Test using Tkinter.
- Testing with Real Life / Noisy Data.
- Attendance in Classrooms.
- Voice Based Security in addition to Biometrics.