Evaluation Distance Measures Between Gaussian Mixture Models of MFCCs

Jensen, Jesper Hojvang; Ellis, Daniel P. W.; Christensen, Mads G.; Jensen, Soren Holdt

In music similarity and in the related task of genre classification, a distance measure between Gaussian mixture models is frequently needed. We present a comparison of the Kullback-Leibler distance, the earth movers distance and the normalized L2 distance for this application. Although the normalized L2 distance was slightly inferior to the Kullback-Leibler distance with respect to classification performance, it has the advantage of obeying the triangle inequality, which allows for efficient searching.


Also Published In

ISMIR 2007: Proceedings of the 8th International Conference on Music Information Retrieval: September 23-27, 2007, Vienna, Austria
Austrian Computer Society

More About This Work

Academic Units
Electrical Engineering
Published Here
June 27, 2012