Open Access Honors Program Thesis
Content retrieval in musical collections has been dependent on textual metadata (e.g. ID3 tags) which can present problems when the title of a piece is forgotten, misspelled, or when the search revolves around the similarity of sound. Content-based MIR (musical information retrieval) could offer an alternative. BLAST (basic local alignment search tool), an algorithm widely used in bioinformatics to search for sequences of aminoacids within longer sequences, seeks similarities and homologies, which makes it interesting for MIR, because musical information can be expected to be imprecise, and because homologies can allow to draw connections between musical pieces. Increased availability of digital music necessitates MIR methods which would allow to search a polyphonic sound collection with polyphonic queries to retrieve individual files, and a question can be raised how viable is BLAST-based retrieval for this kind of data. This paper discusses an implementation of such a system.
Year of Submission
Department of Computer Science
University Honors Designation
A thesis submitted in partial fulfillment of the requirements for the designation University Honors
1 PDF file (21 pages)
©2008 Fedor Aleksandrovich Korsakov
Korsakov, Fedor Aleksandrovich, "Application of Blast-Based Techniques For Musical Information Retrieval" (2008). Honors Program Theses. 800.