Speaker recognition using MFCC and VQ. How to set threshold on minimum VQ distortion?

[Nishat Anjum]

I am doing Speaker Recognition project using MFCC and VQLBG.
In my project, the system identifies speaker with minimum VQ distortion. But I want a threshold for this minimum VQ distortion. That is, if this minimum VQ distortion is above this threshold, the speaker will be considered unregistered. How do i set this threshold? Please help..

 

[iVenky]

I have the code for speaker recognition using MFCC and neural networks. I didn't use VQ for pattern recognition though. I think that neural networks is very effective compared to VQ. If you need the code send private message to me.