AN IMPLEMENTATION METHOD OF A HETEROGENEOUS ASSOCIATIVE MEDIA DATA SEARCH FOR MUSIC DATA AND IMAGE DATA

摘要

This paper presents an implementation method of a heterogeneous associative media data search for music data and image data. It is important to realize a multimedia system which operates and integrates various kinds of multimedia data systematically in the framework of human sensitivity "Kansei." One of the important issues is a realization of media search corresponding to human sensitivity "Kan-sei" with less difficulty for a user. This method is realized using media-lexicon transformation operator, mathematical model of meaning, and Kansei operator. Media-lexicon transformation operator can extract of the impression words which corresponds to impressions of media data. Mathematical model of meaning can measure of the relation of each word extracted from this operator. The Kansei operator is used to adjust the expression with the interpretation of human sensitivity by the logarithm function based on Fechner's law. This method can realize integration between each existing media data only by implementing mechanism to convert into words from media data in each media. The feature of this method is a realization of integration between music data and image data.