Article details

Research area
Biometrics

Location
Pattern Recognition

Date
2002

Author(s)
Ravi P. Ramachandran, Kevin Farrell, Roopashri Ramachandran, Richard J. Mammone

Speaker recognition – general classifier approaches and data fusion methods

Synopsis:

Speaker recognition refers to the concept of recognizing a speaker by his/her voice or speech samples. Some of the important applications of speaker recognition include customer verification for bank transactions, access to bank accounts through telephones, control on the use of credit cards, and for security purposes in the Army, Navy and Air Force. This paper is purely a tutorial that presents a review of the classifier based methods used for speaker recognition. Both unsupervised and supervised classifiers are described. In addition, practical approaches that utilize diversity, redundancy and fusion strategies are discussed with the aim of improving performance.

Read/download now