An online em algorithm in hidden (semi-)Markov models for audio segmentation and clustering

Author

Bietti, Alberto ; Bach, Francis ; Cont, Arshia

Author_Institution

Ircam, UPMC, Paris, France

fYear

2015

fDate

19-24 April 2015

Firstpage

1881

Lastpage

1885

Abstract

Audio segmentation is an essential problem in many audio signal processing tasks, which tries to segment an audio signal into homogeneous chunks. Rather than separately finding change points and computing similarities between segments, we focus on joint segmentation and clustering, using the framework of hidden Markov and semi-Markov models. We introduce a new incremental EM algorithm for hidden Markov models (HMMs) and show that it compares favorably to existing online EM algorithms for HMMs. We present results for real-time segmentation of musical notes and acoustic scenes.

Keywords

acoustic signal processing; audio signal processing; expectation-maximisation algorithm; hidden Markov models; learning (artificial intelligence); music; pattern clustering; audio clustering; audio signal processing tasks; audio signal segmentation; hidden Markov models; online EM algorithm; online learning; real-time acoustic scene segmentation; real-time musical note segmentation; semi Markov models; Acoustics; Approximation algorithms; Clustering algorithms; Hidden Markov models; Real-time systems; Signal processing; Signal processing algorithms; EM algorithm; Hidden Markov models; audio segmentation; online learning;

fLanguage

English

Publisher

ieee

Conference_Titel

Acoustics, Speech and Signal Processing (ICASSP), 2015 IEEE International Conference on

Conference_Location

South Brisbane, QLD

Type

conf

DOI

10.1109/ICASSP.2015.7178297

Filename

7178297