Music genre classification based on dynamical models

Alberto Garćia-Duŕan; Jeŕonimo Arenas-Garćia; Daŕio Garćia-Garćia; Emilio Parrado-Herńandez

Music genre classification based on dynamical models

Alberto Garćia-Duŕan^*, Jeŕonimo Arenas-Garćia, Daŕio Garćia-Garćia, Emilio Parrado-Herńandez

^*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

1 Citation (Scopus)

Abstract

This paper studies several alternatives to extract dynamical features from hidden Markov Models (HMMs) that are meaningful for music genre supervised classification. Songs are modelled using a three scale approach: a first stage of short term (milliseconds) features, followed by two layers of dynamical models: a multivariate AR that provides mid term (seconds) features for each song followed by an HMM stage that captures long term (song) features shared among similar songs. We study from an empirical point of view which features are relevant for the genre classification task. Experiments on a database including pieces of heavy metal, punk, classical and reggae music illustrate the advantages of each set of features.

Original language	English
Title of host publication	ICPRAM 2012 - Proceedings of the 1st International Conference on Pattern Recognition Applications and Methods
Pages	250-256
Number of pages	7
Publication status	Published - 2012
Event	1st International Conference on Pattern Recognition Applications and Methods, ICPRAM 2012 - Vilamoura, Algarve, Portugal Duration: 6 Feb 2012 → 8 Feb 2012

Publication series

Name	ICPRAM 2012 - Proceedings of the 1st International Conference on Pattern Recognition Applications and Methods
Volume	2

Conference

Conference	1st International Conference on Pattern Recognition Applications and Methods, ICPRAM 2012
Country/Territory	Portugal
City	Vilamoura, Algarve
Period	6/02/12 → 8/02/12

Cite this

Garćia-Duŕan, A., Arenas-Garćia, J., Garćia-Garćia, D., & Parrado-Herńandez, E. (2012). Music genre classification based on dynamical models. In ICPRAM 2012 - Proceedings of the 1st International Conference on Pattern Recognition Applications and Methods (pp. 250-256). (ICPRAM 2012 - Proceedings of the 1st International Conference on Pattern Recognition Applications and Methods; Vol. 2).

Garćia-Duŕan, Alberto ; Arenas-Garćia, Jeŕonimo ; Garćia-Garćia, Daŕio et al. / Music genre classification based on dynamical models. ICPRAM 2012 - Proceedings of the 1st International Conference on Pattern Recognition Applications and Methods. 2012. pp. 250-256 (ICPRAM 2012 - Proceedings of the 1st International Conference on Pattern Recognition Applications and Methods).

@inproceedings{619e43975a9644f3a4c6b9d49691495d,

title = "Music genre classification based on dynamical models",

abstract = "This paper studies several alternatives to extract dynamical features from hidden Markov Models (HMMs) that are meaningful for music genre supervised classification. Songs are modelled using a three scale approach: a first stage of short term (milliseconds) features, followed by two layers of dynamical models: a multivariate AR that provides mid term (seconds) features for each song followed by an HMM stage that captures long term (song) features shared among similar songs. We study from an empirical point of view which features are relevant for the genre classification task. Experiments on a database including pieces of heavy metal, punk, classical and reggae music illustrate the advantages of each set of features.",

keywords = "Dynamical features, Genre classification, HMMs, Music retrieval",

author = "Alberto Gar{\'c}ia-Du{\'r}an and Je{\'r}onimo Arenas-Gar{\'c}ia and Da{\'r}io Gar{\'c}ia-Gar{\'c}ia and Emilio Parrado-Her{\'n}andez",

year = "2012",

language = "English",

isbn = "9789898425980",

series = "ICPRAM 2012 - Proceedings of the 1st International Conference on Pattern Recognition Applications and Methods",

pages = "250--256",

booktitle = "ICPRAM 2012 - Proceedings of the 1st International Conference on Pattern Recognition Applications and Methods",

note = "1st International Conference on Pattern Recognition Applications and Methods, ICPRAM 2012 ; Conference date: 06-02-2012 Through 08-02-2012",

}

Garćia-Duŕan, A, Arenas-Garćia, J, Garćia-Garćia, D & Parrado-Herńandez, E 2012, Music genre classification based on dynamical models. in ICPRAM 2012 - Proceedings of the 1st International Conference on Pattern Recognition Applications and Methods. ICPRAM 2012 - Proceedings of the 1st International Conference on Pattern Recognition Applications and Methods, vol. 2, pp. 250-256, 1st International Conference on Pattern Recognition Applications and Methods, ICPRAM 2012, Vilamoura, Algarve, Portugal, 6/02/12.

Music genre classification based on dynamical models. / Garćia-Duŕan, Alberto; Arenas-Garćia, Jeŕonimo; Garćia-Garćia, Daŕio et al.
ICPRAM 2012 - Proceedings of the 1st International Conference on Pattern Recognition Applications and Methods. 2012. p. 250-256 (ICPRAM 2012 - Proceedings of the 1st International Conference on Pattern Recognition Applications and Methods; Vol. 2).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Music genre classification based on dynamical models

AU - Garćia-Duŕan, Alberto

AU - Arenas-Garćia, Jeŕonimo

AU - Garćia-Garćia, Daŕio

AU - Parrado-Herńandez, Emilio

PY - 2012

Y1 - 2012

N2 - This paper studies several alternatives to extract dynamical features from hidden Markov Models (HMMs) that are meaningful for music genre supervised classification. Songs are modelled using a three scale approach: a first stage of short term (milliseconds) features, followed by two layers of dynamical models: a multivariate AR that provides mid term (seconds) features for each song followed by an HMM stage that captures long term (song) features shared among similar songs. We study from an empirical point of view which features are relevant for the genre classification task. Experiments on a database including pieces of heavy metal, punk, classical and reggae music illustrate the advantages of each set of features.

AB - This paper studies several alternatives to extract dynamical features from hidden Markov Models (HMMs) that are meaningful for music genre supervised classification. Songs are modelled using a three scale approach: a first stage of short term (milliseconds) features, followed by two layers of dynamical models: a multivariate AR that provides mid term (seconds) features for each song followed by an HMM stage that captures long term (song) features shared among similar songs. We study from an empirical point of view which features are relevant for the genre classification task. Experiments on a database including pieces of heavy metal, punk, classical and reggae music illustrate the advantages of each set of features.

KW - Dynamical features

KW - Genre classification

KW - HMMs

KW - Music retrieval

UR - http://www.scopus.com/inward/record.url?scp=84862187253&partnerID=8YFLogxK

M3 - Conference contribution

SN - 9789898425980

T3 - ICPRAM 2012 - Proceedings of the 1st International Conference on Pattern Recognition Applications and Methods

SP - 250

EP - 256

BT - ICPRAM 2012 - Proceedings of the 1st International Conference on Pattern Recognition Applications and Methods

T2 - 1st International Conference on Pattern Recognition Applications and Methods, ICPRAM 2012

Y2 - 6 February 2012 through 8 February 2012

ER -

Garćia-Duŕan A, Arenas-Garćia J, Garćia-Garćia D, Parrado-Herńandez E. Music genre classification based on dynamical models. In ICPRAM 2012 - Proceedings of the 1st International Conference on Pattern Recognition Applications and Methods. 2012. p. 250-256. (ICPRAM 2012 - Proceedings of the 1st International Conference on Pattern Recognition Applications and Methods).

Music genre classification based on dynamical models

Abstract

Publication series

Conference

Other files and links

Fingerprint

Cite this