everybody i in a single button so that a new feature for automatic speaker verification
spoofing
the data for us to show that automatic specific speaker verification system
are all the spoofing
so
especially when no we are in an ideal scenario where the spoofing we cannot be
you know in advance so the need the all the for the generalize the going
to measure
so we propose an feature example sample transform
and the t v is a
task automation
and different from a for their trust one because we suppose that fourier transform me
like frequency resolution secrecy security uses our body a time frequency the solution that means
i your time resolution for i frequencies and the higher frequency resolution for lower frequencies
we combine the c d more or less with the traditional a cepstral analysis but
we found the problem but using the discrete cosine transform or you know the identity
applied before so
for two reasons results secrecy a dct have different skate one dramatic another one is
the dinner and dramatically dct basis are no longer than one
so we found a total shock joe's the only for me to sing but they're
not but for thirty twenty scale over the speaker i
with a linear frequency scale
a
these is the some form comparison of a results a nice peaceful for database and
we know that
we found that no i x from a tax or the system and earlier are
excellent error rate for unknown i so the thing to me
best performance especially for a s then that young federation seem to use these your
voice
and where a we obtain a seven percent over relative improvement
embodied the exposing detection performance
a part of to be in our work seeks
we that seventy two relative improvement
so for more details are wait for you
sure some possible