A light-weight concept ontology for annotating digital music.

In the recent time, the digital music items on the internet have been evolving to an enormous information space where we try to find/locate the piece of information of our choice by means of search engine. The current trend of searching for music by means of music consumers' keywords/tags is unable to provide satisfactory search results; and search and retrieval of music may be potentially improved if music metadata is created from semantic information provided by association of end-users' tags with acoustic metadata which is easy to extract automatically from digital music items. Based on this observation, our research objective was to investigate how music producers may be able to annotate music against MPEG-7 description (with its acoustic metadata) to deliver meaningful search results. In addressing this question, we investigated the potential of multimedia ontologies to serve as backbone for annotating music items and prospective application scenarios of semantic technologies in the digital music industry. We achieved with our main contribution under this thesis is the first prototype of mpeg-7Music annotation ontology that establishes a mapping of end-users tags with MPEG-7 acoustic metadata as well as extends upper level multimedia ontologies with end-user tags. Additionally, we have developed a semi-automatic annotation tool to demonstrate the potential of the mpeg-7Music ontology to serve as light weight concept ontology for annotating digital music by music producers. The proposed ontology has been encoded in dominant semantic web ontology standard OWL1.0 and provides a standard interoperable representation of the generated semantic metadata. Our innovations in designing the semantic annotation tool were focussed on supporting the music annotation vocabulary (i.e. the mpeg-7Music) in an attempt to turn the music metadata information space to a knowledgebase.