search

ALIZE-Speaker-Recognition / LIA_RAL

A high-level toolkit for speaker recognition, build on top of ALIZE-Core.
http://alize.univ-avignon.fr
GNU Lesser General Public License v3.0
125 stars 27 forks source link

int decision=0; in ComputeNorm.cpp!! #28

Open MSAlghamdi opened 5 years ago

MSAlghamdi commented 5 years ago

In ComputeNorm, decision=0 and never updated dynamically to a value that represent the result. It's defined as zero variable and stays zero all the way till it's printed zero in the result.

How can we let it one for highest score and zero for the rest?!

jfb84 commented 5 years ago

Hi That's true... It was designed for NIST evaluation where we used a scoring script to put the decision after all the processes (as we had usually several systems +fusion). JF

De: "MSAlghamdi" notifications@github.com À: "ALIZE-Speaker-Recognition" LIA_RAL@noreply.github.com Cc: "Subscribed" subscribed@noreply.github.com Envoyé: Lundi 3 Décembre 2018 04:23:07 Objet: [ALIZE-Speaker-Recognition/LIA_RAL] int decision=0; in ComputeNorm.cpp!! (#28)

In ComputeNorm, decision=0 and never updated dynamically to a value that represent the result. It's defined as zero variable and stays zero all the way till it's printed zero in the result.

How can we let it one for highest score and zero for the rest?!

— You are receiving this because you are subscribed to this thread. Reply to this email directly, [ https://github.com/ALIZE-Speaker-Recognition/LIA_RAL/issues/28 | view it on GitHub ] , or [ https://github.com/notifications/unsubscribe-auth/ATmd7034FaLASQZzoYY-hv6ji_IBmaafks5u1JkbgaJpZM4Y9z0d | mute the thread ] .

--

Jean-Francois BONASTRE Directeur du LIA LIA/CERI Université d'Avignon Tel: +33/0 490843514 directeur-lia@univ-avignon.fr @jfbonastre

MSAlghamdi commented 5 years ago

Could you please refer me to the scoring script you mentioned?

jfb84 commented 5 years ago

not sure we still have it (me no) as it we updated it each year due to NIST-SRE file format changes. Basically, it was reading the output file line by line and if the LLR was higher than the threshold put 1 in hte decision field, 0 if not... Easy to reimplement that as an option in ComputeNorm if you need it (taking the example of ComputeTest which have this option)

De: "MSAlghamdi" notifications@github.com À: "ALIZE-Speaker-Recognition" LIA_RAL@noreply.github.com Cc: "Jean-Francois Bonastre" jean-francois.bonastre@univ-avignon.fr, "Comment" comment@noreply.github.com Envoyé: Lundi 3 Décembre 2018 08:03:19 Objet: Re: [ALIZE-Speaker-Recognition/LIA_RAL] int decision=0; in ComputeNorm.cpp!! (#28)

Could you please refer me to the scoring script you mentioned?

— You are receiving this because you commented. Reply to this email directly, [ https://github.com/ALIZE-Speaker-Recognition/LIA_RAL/issues/28#issuecomment-443608829 | view it on GitHub ] , or [ https://github.com/notifications/unsubscribe-auth/ATmd71LqeeVUGrrh55WUpPCW8vzZTcvBks5u1My3gaJpZM4Y9z0d | mute the thread ] .

--

Jean-Francois BONASTRE Directeur du LIA LIA/CERI Université d'Avignon Tel: +33/0 490843514 directeur-lia@univ-avignon.fr @jfbonastre