Research on the integrated method of the associative search engine systems for different-species media data

本研究通过元数据基础系统集作为元层次来实现各不同参考系统的集成方法。该方法可以在不改变现有参照系组成的前提下，整合现有参照系。该方法通过在元级上仅实现原语操作符，将现有参考系统集成为关联搜索引擎中的新功能。该方法通过在元级别上仅新实现原语操作符，使现有参考系统的集成能够在关联搜索引擎中合并新功能。本研究实现了将语义关联搜索与模式匹配相结合的媒体信息系统，并通过实验结果验证了该方法的有效性。模式匹配通过重复数据的简单排序操作进行过滤。语义关联搜索获取具有相近含义的信息，并将数据按照与用户印象最相关的顺序重新排列。通过对各参考系统的集成，实现了两种功能的高效媒体信息检索。作为未来的部署，提出了一种不同类型媒体数据集成方法的实现；异构媒体到媒体搜索，例如，从一张图片中整合合适的音乐。

This research realized the method with integration of each different reference systems through the metadata base system set as the meta-level. This method can integrate the existing reference systems without changing the composition of these. This method enables integration of existing reference systems to incorporate as a new function in an associative search engine by newly realizing only the primitive operator on a meta-level. This method enables integration of existing reference systems to incorporate new functions as these in an associative search engine by newly realizing only the primitive operator on a meta-level.In this research, the media information system which integrated semantic associative search with pattern matching was realized, and we clarify the effectiveness of this method by showing experimental results. The pattern matching filters by repetition of the simple collation operation of data. The semantic associative search acquires information with a near meaning, and this method can rearrange data in order of nearest correlation with a user's impression. We realized highly efficient media information retrieval with both of functions by integration of each reference system.As future deployment, realization of the integrated method of media data of a different kind is mentioned ; heterogeneous media to media search, e.g., integrating appropriate music from a picture.