图像数据库聚类索引方法研究  

Cluster-based Index Method for Image Database

作　　者：龙巧云[1] 邓彬[1] 曾敏[1] 

机构地区：[1]湖南工业大学信息系,湖南株洲412000

出　　处：《武汉理工大学学报（信息与管理工程版）》2008年第4期541-543,共3页Journal of Wuhan University of Technology：Information & Management Engineering

摘　　要：针对大规模图像数据库的特点,用基于高斯混合模型的期望值最大(EM)算法对图像数据库进行聚类划分,为每个聚类建立索引项,且它所包含的原始特征数据在磁盘上连续存储。查询时,对于用户的查询范例首先确定最可能的候选聚类,然后在候选聚类范围内查询翔实图像。试验结果表明,该方法可提高图像检索速度。According to the characteristics of the large-scale image database, the image database was classified by a study of EM learning algorithm based on Gaussian Mixture Model.The clusters here are called index clusters and the original feature vectors of the image are stored continuously. For a given query ,the first phase computes the distances between the query example and each cluster index and returns the clusters with the smallest distance, here namely candidate clusters;then the second phase retrieves the original feature vectors within the candidate clusters to gain the approximate nearest neighbors. The proposed method can soeed uo the index searching.

关 键 词：图像库 聚类 图像检索 高斯混合模型 EM算法 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]