Designing a Music Content Retrieval System supporting Korean

Abstract

In recent years, as increase of music content market, Various studies applying the user's emotional expressions to search for content have been performed. Especially, research on emotion model and of ontology is representative. Those results have been applied in MIR (Music Information Retrieval) system. However, an ability of searching and classifying the content by expressions of emotion is somewhat lacking in MIR system and the kind of words for using in search is simple yet. In addition, it is difficult to process search for content using words of emotion expressed in Korean because Emotion model was designed for English-based. Therefore, in this paper, propose the expanded Emotion model (XEM) for supporting Korean, and define the ontology (MEMO) based on XEM. Through them, rich and accurate search for music content using the vocabulary expressed in Korean becomes available.