@inproceedings{7fa5ebc19fde4b0b9f30003196e884d1,
title = "Fusion of acoustic and tokenization features for speaker recognition",
abstract = "This paper describes our recent efforts in exploring effective discriminative features for speaker recognition. Recent researches have indicated that the appropriate fusion of features is critical to improve the performance of speaker recognition system. In this paper we describe our approaches for the NIST 2006 Speaker Recognition Evaluation. Our system integrated the cepstral GMM modeling, cepstral SVM modeling and tokenization at both phone level and frame level. The experimental results on both NIST 2005 SRE corpus and NIST 2006 SRE corpus are presented. The fused system achieved 8.14\% equal error rate on 1conv4w-1conv4w test condition of the NIST 2006 SRE.",
keywords = "Cepstral feature, Gaussian mixture model, Phonotactic feature, Speaker recognition, Support vector machine, Tokenization, fusion",
author = "Rong Tong and Bin Ma and Lee, \{Kong Aik\} and Changhuai You and Donglai Zhu and Tomi Kinnunen and Hanwu Sun and Minghui Dong and Chng, \{Eng Siong\} and Haizhou Li",
year = "2006",
doi = "10.1007/11939993\_59",
language = "English",
isbn = "3540496653",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
pages = "566--577",
booktitle = "Chinese Spoken Language Processing - 5th International Symposium, ISCSLP 2006, Proceedings",
note = "5th International Symposium on Chinese Spoken Language Processing, ISCSLP 2006 ; Conference date: 13-12-2006 Through 16-12-2006",
}