time
so we first moved with allocation problem gmm-ubm systems
the motivation is the biometric characteristics are not secret in the h increasing surveillance so
think about you have a financial institution use all your one data there
not the financial situation that was we line whose line and biometrics need to be
projected on this protection it's be strong and from
also to financial institutions got take that shouldn't be a lot that from one
the data from one you cannot apply to the other issue transfer there
just as an example and that's why we want to real research but what's biometric
template protection for speaker recognition
where you can go for biometric crypto system or can similar biometrics
we wanted to
emphasized on the properties of the ice was extended to four seven four five
which is about biometric template protection and the basic things on there are very simple
energy so i got the biometric one the data and i cannot which we've
any thing right because it's to do much to not
and unlike ability so i five and one system i cannot retrieve any information one
of the system although they are based on the same ubm for example
and what we propose a small bit allocation
you can use what follows for the commitment or for divorce or schemes also for
fast searches and basic idea is to extract common mfcc features and that supervector and
then combined voice the supervectors
and four button to template protection scheme using for example haven't distance comparison
for improving the performance we also we did expected the most discriminant features so for
that we use the relevant projection thank can see that post