A sampling-based speaker clustering using utterance-oriented Dirichlet process mixture model and its evaluation on large-scale data. (28th October 2015)

• Record Type: Journal Article
• Authors: Tawara, Naohiro; Ogawa, Tetsuji; Watanabe, Shinji; Nakamura, Atsushi; Kobayashi, Tetsunori
• Is Part Of: APSIPA transactions on signal and information processing. Volume 4(2015)
• Publication Date: 2015-10-28
• Subjects: Sampling approach, -- Non-parametric Bayesian model, -- Gibbs sampling, -- Utterance-oriented Dirichlet process mixture model, -- Speaker clustering; Signal processing -- Periodicals; 621.3822
• DOI: 10.1017/ATSIP.2015.19
• Languages: English
• ISSNs: 2048-7703

Abstract

Abstract : An infinite mixture model is applied to model-based speaker clustering with sampling-based optimization to make it possible to estimate the number of speakers. For this purpose, a framework of non-parametric Bayesian modeling is implemented with the Markov chain Monte Carlo and incorporated in the utterance-oriented speaker model. The proposed model is called the utterance-oriented Dirichlet process mixture model (UO-DPMM). The present paper demonstrates that UO-DPMM is successfully applied on large-scale data and outperforms the conventional hierarchical agglomerative clustering, especially for large amounts of utterances.