A design of a Kansei retrieval system for distributed multimedia databases

Abstract

We designed a Kansei retrieval system, which facilitates integrated retrieval of multimedia data from several distributed multimedia databases on the network according to user's sensibility. This system is based on the method we proposed, of semi-automatic indexing with emotional keywords such as genre names for multimedia database (e.g. movie files, audio files) according to user's sensibility by using users' access histories for databases. In order to facilitate integrated Kansei retrieval, we designed the system with a meta-search engine, and also designed the protocol among clients, meta-search engine, and databases.