Title :
Maximizing Distance between GMMs for Speaker Verification
Author :
Kim, Min-Seok ; Yang, Il-Ho ; Yu, Ha-Jin
Author_Institution :
Sch. of Comput. Sci., Univ. of Seoul, Seoul
Abstract :
In this paper, we propose a feature transformation method to maximize the distances between the Gaussian mixture models for speaker verification. The feature transformation matrix is optimized by using particle swarm optimization. We evaluate the transformation using YOHO speech data, and the transformation is applied to some speakers who give poor performance. As the result, the overall equal error rate is reduced to 1.71% from 1.97% of the baseline.
Keywords :
Gaussian processes; matrix algebra; particle swarm optimisation; speaker recognition; GMM; Gaussian mixture models; YOHO speech data; feature transformation matrix; feature transformation method; particle swarm optimization; speaker verification; Computer science; Covariance matrix; Density functional theory; Error analysis; Feature extraction; Humans; Optimization methods; Particle swarm optimization; Speaker recognition; Speech analysis; GMM; PSO; Speaker Verification;
Conference_Titel :
Natural Computation, 2008. ICNC '08. Fourth International Conference on
Conference_Location :
Jinan
Print_ISBN :
978-0-7695-3304-9
DOI :
10.1109/ICNC.2008.820
