On music genre classification via compressive sampling

Publication: Research - peer-reviewConference article in Journal

Abstract

Recent work \cite{Chang2010} combines low-level acoustic features
and random projection (referred to as ``compressed sensing'' in \cite{Chang2010})
to create a music genre classification system showing
an accuracy among the highest reported
for a benchmark dataset.
This not only contradicts previous findings
that suggest low-level features are inadequate for
addressing high-level musical problems,
but also that a random projection of features
can improve classification.
We reproduce this work and resolve these contradictions.
Close

Details

Recent work \cite{Chang2010} combines low-level acoustic features
and random projection (referred to as ``compressed sensing'' in \cite{Chang2010})
to create a music genre classification system showing
an accuracy among the highest reported
for a benchmark dataset.
This not only contradicts previous findings
that suggest low-level features are inadequate for
addressing high-level musical problems,
but also that a random projection of features
can improve classification.
We reproduce this work and resolve these contradictions.
Original languageEnglish
JournalInternational Conference on Multimedia and Expo
StatePublished - 2013
Event2012 IEEE Conference on Multimedia & Expo - San Jose, United States

Conference

Conference2012 IEEE Conference on Multimedia & Expo
CountryUnited States
CitySan Jose
Period15/07/201319/07/2013

Download statistics

No data available
ID: 71871680