A melody-based similarity computation algorithm for musical information

A melody-based similarity computation algorithm for musical information is proposed in which melodies are used as the basic features for calculating the similarity between two music files. In multimedia information systems, in order to retrieve the needed information appropriately, content-based retrieval strategies are used which are based on the similarity between the user´s query and the stored multimedia information. Hence, an effective similarity computation algorithm is basically required for content-based retrieval for music information. In the proposed algorithm, the pitches and lengths of the musical notes which compose the melodies of two corresponding music files are compared, and then the comparison results are finally condensed into a similarity value between the two corresponding music files. Experimental results show that the proposed algorithm is able to calculate the similarity value between music files appropriately on the basis of melodies