To access the full text documents, please follow this link:

Deep belief networks for i-vector based speaker recognition
Ghahabi Esfahani, Omid; Hernando Pericás, Francisco Javier
Universitat Politècnica de Catalunya. Departament de Teoria del Senyal i Comunicacions; Universitat Politècnica de Catalunya. VEU - Grup de Tractament de la Parla
The use of Deep Belief Networks (DBNs) is proposed in this paper to model discriminatively target and impostor i-vectors in a speaker verification task. The authors propose to adapt the network parameters of each speaker from a background model, which will be referred to as Universal DBN (UDBN). It is also suggested to backpropagate class errors up to only one layer for few iterations before to train the network. Additionally, an impostor selection method is introduced which helps the DBN to outperform the cosine distance classifier. The evaluation is performed on the core test condition of the NIST SRE 2006 corpora, and it is shown that 10% and 8% relative improvements of EER and minDCF can be achieved, respectively.
Peer Reviewed
Àrees temàtiques de la UPC::Enginyeria de la telecomunicació
Àrees temàtiques de la UPC::Informàtica
Neural networks (Computer science)
Deep belief network
Neural network
Speaker recognition
Xarxes neuronals (Informàtica)
Institute of Electrical and Electronics Engineers (IEEE)

Show full item record

Related documents

Other documents of the same author

Ghahabi Esfahani, Omid; Hernando Pericás, Francisco Javier
Ghahabi Esfahani, Omid; Hernando Pericás, Francisco Javier
Hernando Pericás, Francisco Javier; Hernando Pericás, Francisco Javier
Zelenak, Martin; Schulz, Henrik; Hernando Pericás, Francisco Javier
Safari, Pooyan; Ghahabi, Omid; Hernando Pericás, Francisco Javier