Towards incorporating the notion of feature shape in music and text retrieval

Cheng, Yi-Yun; Weigl, David; Downie, J. Stephen; Page, Kevin

Permalink

https://hdl.handle.net/2142/100247 Copy

Description

Title

Towards incorporating the notion of feature shape in music and text retrieval

Author(s)

• Cheng, Yi-Yun
• Weigl, David
• Downie, J. Stephen
• Page, Kevin

Issue Date

2018

Keyword(s)

• feature data
• feature shape
• music information retrieval
• HathiTrust

Date of Ingest

2018-07-12T15:37:16Z

Abstract

Extracted feature data augment information resources with concrete characterizations of their content, but only approximate to the meaningful high-level descriptions typically expected by digital musicology scholars (domain experts with some technological affinity, but with no expertise in signal processing or feature data). Feature shapes provide abstract aggregations of feature types which share common characteristics when applied in extraction workflows. We explore the feasibility of feature shape-based filtering and querying within a large audio dataset of live music performances, employing operation sequences as specified by the Audio Feature Ontology and Vocabulary. We further implement analogous semantic structures for the HathiTrust Extracted Feature Dataset to demonstrate the general applicability of feature shapes in music and text retrieval.

Publisher

iSchools

Series/Report Name or Number

iConference 2018 Proceedings

Type of Resource

text

Genre of Resource

Conference Poster

Language

eng

Permalink

http://hdl.handle.net/2142/100247

Copyright and License Information

Copyright 2018 is held by Yi-Yun Cheng, David Weigl, J. Stephen Downie, Kevin Page. Copyright permissions, when appropriate, must be obtained directly from the authors.

Owning Collections