Spectrum histograms and fluctuation patterns are representations of audio fragments. By comparing these representations, we can determine the similarity between the corresponding fragments. Traditionally, this is done using the Euclidean distance. In this chapter, however, we study an alternative approach, namely, comparing the representations by means of fuzzy similarity measures. Once the preliminary notions have been addressed, we present a recently introduced triparametric family of fuzzy similarity measures, together with several constraints on its parameters that warrant certain potentially desirable or useful properties. In particular, we present constraints for several forms of restrictability, which allow to reduce the computation time in practical applications. Next, we use some members of this family to construct various audio similarity measures based on spectrum histograms and fluctuation patterns. To conclude, we analyse the performance of the constructed audio similarity measures experimentally. © 2008 Springer-Verlag Berlin Heidelberg.
CITATION STYLE
Bosteels, K., & Kerre, E. E. (2008). Fuzzy audio similarity measures based on spectrum histograms and fluctuation patterns. Studies in Computational Intelligence, 96, 213–231. https://doi.org/10.1007/978-3-540-76827-2_9
Mendeley helps you to discover research relevant for your work.