Skip to main navigation Skip to search Skip to main content

Large vocabulary automatic chord estimation with an even chance training scheme

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

18 Citations (Scopus)

Abstract

This paper presents a large vocabulary automatic chord estimation system implemented using a bidirectional long short-term memory recurrent neural network trained with a skewed-class-aware scheme. This scheme gives the uncommon chord types much more exposure during the training process. The evaluation results indicate that: compared with a normal training scheme, the proposed scheme can boost the weighted chord symbol recalls of some uncommon chords and significantly improve the average chord quality accuracy, at the expense of the overall weighted chord symbol recall.

Original languageEnglish
Title of host publicationProceedings of the 18th International Society for Music Information Retrieval Conference, ISMIR 2017
EditorsSally Jo Cunningham, Zhiyao Duan, Xiao Hu, Douglas Turnbull
Pages531-536
Number of pages6
ISBN (Electronic)9789811151798
Publication statusPublished - 2017
Externally publishedYes
Event18th International Society for Music Information Retrieval Conference, ISMIR 2017 - Suzhou, China
Duration: 23 Oct 201727 Oct 2017

Publication series

NameProceedings of the 18th International Society for Music Information Retrieval Conference, ISMIR 2017

Conference

Conference18th International Society for Music Information Retrieval Conference, ISMIR 2017
Country/TerritoryChina
CitySuzhou
Period23/10/1727/10/17

Fingerprint

Dive into the research topics of 'Large vocabulary automatic chord estimation with an even chance training scheme'. Together they form a unique fingerprint.

Cite this