SIA(M)ESE: An Algorithm for Transposition Invariant, Polyphonic Content-Based Music Retrieval

We introduce a novel algorithm for transposition-invariant contentbased polyphonic music retrieval. Our SIA(M)ESE algorithm is capable of finding transposition invariant occurrences of a given template, in a database of polyphonic music called a dataset. We allow arbitrary gapping, i.e., between musical events in the dataset that have been found to match points in the template, there may be any finite number of other intervening events. SIA(M)ESE can be implemented so that it finds all transposition-invariant complete matches for a -dimensional template of size in a dimensional dataset of size in a worst-case running time of ; another implementation finds even the incomplete matches in time. The algorithm is generalizable to any arbitrary, multidimensional translation invariant pattern matching problem, where the events are representable by points in a multidimensional dataset.