基于多媒体图像数据的查询检索

Retrieve for Multimedia-Based Image Data

本文对多媒体图像数据提出一套较灵活的存储、索引查询方式。针对多媒体数据库信息查询通用性的问题,构建了一个较为集成的特征库。针对高维数据查询效率问题,将基于内容、颜色、距离和形状的匹配算法与FNN算法及查找聚类重心的加速检索算法集成于查询检索引擎IRE,使查询更高效、应用更广泛。

In this paper,a series of storage and retrieval strategies are put forward for multimedia image data. To the problem of general capability of information retrieval for multimedia database,a more integrated feature-base is constructed;to the problem of retrieval effciency of high-dimensional data,some algorithms,such as the match algorithms based on content,color,distance and shape,are integrated with the algorithm FNN and the algorithm of quickening retrieval by finding cluster centers,all these are combined into IRE and make the application of our query method more efficient and broad.