search

Abhay0899193 / Speaker-Recognition

Speaker Recognition System using MFCC and GMM.
24 stars 12 forks source link

Not Recognizing The Unknown Person other than trained people. #2

Closed SanatMoharir closed 4 years ago

SanatMoharir commented 5 years ago

@Abhay0899193 Hey mate, As I tried your code It is working fine after some dodgy changes. But I have a problem what if I want to identify unknown person Ex. right now it is recognizing an untrained person as one of the trained people. I just want untrained or unknown person as an anonymous person. Can you please help me with this issue?

Abhay0899193 commented 5 years ago

Hi

For recognising a unknown person we have to make universal background model which is not implement in the model due to lack of data.

Thank you

On Tue 5 Nov, 2019, 1:26 PM Sanat Moharir <notifications@github.com wrote:

@Abhay0899193 https://github.com/Abhay0899193 Hey mate, As I tried your code It is working fine after some dodgy changes. But I have a problem what if I want to identify unknown person Ex. right now it is recognizing an untrained person as one of the trained people. I just want untrained or unknown person as an anonymous person. Can you please help me with this issue?

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/Abhay0899193/Speaker-Recognition/issues/2?email_source=notifications&email_token=AHIWU4KJAEKQIO7Z7J7OWR3QSERJPA5CNFSM4JI66FH2YY3PNVWWK3TUL52HS4DFUVEXG43VMWVGG33NNVSW45C7NFSM4HW2ATXA, or unsubscribe https://github.com/notifications/unsubscribe-auth/AHIWU4PZH4FQMGU3TWD4JNTQSERJPANCNFSM4JI66FHQ .

SanatMoharir commented 5 years ago

Hi, @Abhay0899193 thanks for the reply! Can you please elaborate on this "Universal Model" I am quite a newbie in this. And how to do this? How to train the universal model and all?

Abhay0899193 commented 5 years ago

background model is basically GMM with a huge data set. This paper explains it better...

On Tue 5 Nov, 2019, 2:22 PM Sanat Moharir <notifications@github.com wrote:

Hi, @Abhay0899193 https://github.com/Abhay0899193 thanks for the reply! Can you please elaborate on this "Universal Model" I am quite a newbie in this. And how to do this? How to train the universal model and all?

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/Abhay0899193/Speaker-Recognition/issues/2?email_source=notifications&email_token=AHIWU4O6DC3CY5FCPS3B7GLQSEX3FA5CNFSM4JI66FH2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEDCCCXQ#issuecomment-549724510, or unsubscribe https://github.com/notifications/unsubscribe-auth/AHIWU4NK4D6A5ZZYHUTWGODQSEX3FANCNFSM4JI66FHQ .

Abhay0899193 commented 5 years ago

https://www.google.com/url?sa=t&source=web&rct=j&url=https://ieeexplore.ieee.org/document/5713236/&ved=2ahUKEwjX86TghdPlAhWTeisKHYXoDAQQFjACegQIBhAB&usg=AOvVaw1OdpTdG15SS0bx-toQYXYA&cshid=1572956064139