5th European Conference on Speech Communication and Technology

Rhodes, Greece
September 22-25, 1997

Speaker Models Designed from Complete Data Sets: A New Approach to Text-Independent Speaker Verification

Dominik R. Dersch (1), Robin W. King (2)

(1) Speech Technology Research Group Department of Electrical Engineering, The University of Sydney, Australia (2) Faculty of Information Technology, University of South Australia, Australia

In this paper we present a new approach to text independent speaker verification. Speaker models are created from complete data sets, derived from a set of sentences. A decision on an identity claim is based on the calculation of the mean next neighbour distance between a speaker model and a test utterance. A Vector quantization technique serves to efficiently extract this frame based similarity measure. It is the purpose of this paper to investigate this new approach and test its performance on a large database as a function of a number of parameters, i.e., the number of data vectors in each model and the length of the test utterance. The best results on a set of 108 speakers are 0:93% false rejection rate and 0:98% false acceptance rate.

Full Paper

Bibliographic reference.  Dersch, Dominik R. / King, Robin W. (1997): "Speaker models designed from complete data sets: a new approach to text-independent speaker verification", In EUROSPEECH-1997, 2283-2286.