instituto de matemáticas universidad de sevilla
Antonio de Castro Brzezicki
Clustering Algorithms in Music Information Retrieval. Application to Flamenco Music
Given growing importance of large digital music libraries, research in Music Information Retrieval (MIR) has recently focused on algorithms for automatic indexing, classification and analysis of audio files. Clustering discipline is related to interesting problems in mathematics. In particular, clustering musical entities into groups of similar objects is considered a major challenge in the field of Music and Mathematics with possible applications in a variety of related tasks, such as genre classification or unsupervised discovery of repeated melodic patterns. In this talk, we present a number of clustering approaches for the specific case of melodic sequences in Flamenco singing and their application in MIR tasks.