Speaker Verification

We have developed an Automatic Speaker Verification system using advanced signal processing and mathematical techniques which can be used to automatically find patterns, which are unique to a person, from the given speech signal. This pattern is stored in the database which will be later used to validate a persons identity.
Each speaker has to utter the password that is assigned to him. When the password is uttered the Speaker Verification program validates the password and the identity of the speaker by performing pattern matching operations. If the operation succeeds then a message "Correct" is displayed. If the speaker tries to imitate another speakers password then the Speaker Verification program displays a message "Incorrect" (In this demo the password is the name of the speaker). The Speaker Verification application has a false acceptance rate close to 0%.The entire task of Speaker Verification is done in real time, as shown in the video.

Category:

Howto & Style

Tags:

• Speaker
• Verification
• Biometrics
• Security
• Signal
• Processing

[cigp]

if you say that the false acceptance rate (f.a.r) is close to 0% , you must say how many individuals has the database. i.e 0% f.a.r in a 200 individual test is better than 0% f.a.r in a 5 individual test.

has you any publication reference for this work

[Riyanat]

I need to develop a JAVA application like this for my final year project. How did you go about it pls??

[ainsleymg]

how do you get it