Using Duration Models To Reduce Fragmentation in Audio Segmentation

Abdallah, Samer; Sandler, Mark; Rhodes, Christophe and Casey, Michael A.. 2006. Using Duration Models To Reduce Fragmentation in Audio Segmentation. Machine Learning, 65(2-3), pp. 485-515. ISSN 0885-6125 [Article]

No full text available

Abstract or Description

We investigate explicit segment duration models in addressing the problem of fragmentation in musical audio segmentation. The resulting probabilistic models are optimised using Markov Chain Monte Carlo methods; in particular, we introduce a modification to Wolff's algorithm to make it applicable to a segment classification model with an arbitrary duration prior. We apply this to a collection of pop songs, and show experimentally that the generated segmentations suffer much less from fragmentation than those produced by segmentation algorithms based on clustering, and are closer to an expert listener's annotations, as evaluated by two different performance measures.

Item Type:

Article

Identification Number (DOI):

https://doi.org/10.1007/s10994-006-0586-4

Related URLs:

Departments, Centres and Research Units:

Computing

Dates:

DateEvent
October 2006Published

Item ID:

1003

Date Deposited:

12 Mar 2009 15:41

Last Modified:

20 Jun 2017 11:52

Peer Reviewed:

Yes, this version has been peer-reviewed.

URI:

https://research.gold.ac.uk/id/eprint/1003

Edit Record Edit Record (login required)