2007:Audio Genre Classification Results

Introduction

These are the results for the 2007 running of the Audio Genre Classification task. For background information about this task set please refer to the 2007:Audio Genre Classification page. The data set consisted of 7000 30 second clips covering the following genres (700 tracks per genre):

A  : BAROQUE
B  : BLUES
C  : CLASSICAL
D  : COUNTRY
E  : EDANCE
F  : JAZZ
G  : METAL
H  : RAPHIPHOP
I  : ROCKROLL
J  : ROMANTIC

Partial scores were earned for some confusions (i.e., ROCKROLL-METAL; JAZZ-BLUES, etc.)

General Legend

Team ID

ME = Michael I. Mandel, Daniel P. W. Ellis
TL = Thomas Lidy, Andreas Rauber, Antonio Pertusa, José Manuel Iñesta
GT = George Tzanetakis
GH = Enric Guaus, Perfecto Herrera
IM = IMIRSEL M2K

Overall Summary Results

MIREX 2007 Audio Genre Classification Summary Results - Raw and Hierarchical Classification Accuracy Averaged Over Three Train/Test Folds

Participant	Avg. Hierarchical Classification Accuracy	Avg. Raw Classification Accuracy
GH	71.87%	62.89%
IM_knn	64.83%	54.87%
IM_svm	76.56%	68.29%
TL	75.57%	66.71%
ME	75.03%	66.60%
ME_spec	73.57%	65.50%
GT	74.15%	65.34%

download these results as csv

MIREX 2007 Audio Genre Classification Evaluation Logs and Confusion Matrices

GH
IM_knn
IM_svm
TL
ME
ME_spec
GT

MIREX 2007 Audio Genre Classification Run Times

Participant	Runtime (sec) / Fold
GH	Feat Ex: 22740 Train/Classify: 194
IM-knn	Feat Ex: 6879 Train/Classify: 1245
IM-svm	Feat Ex: 6879 Train/Classify: 51
TL	Feat Ex: 54192 Train/Classify: 147
ME	Feat Ex: 8166 Train/Classify: 207
ME_spec	Feat Ex: 8018 Train/Classify: 210
GT	Feat Ex/Train/Classify: 1442

download these results as csv