基于细微特征区分的海量图像检索模型仿真  

Simulation of Massive Image Retrieval Model Based on Fine Feature

作　　者：曹锋[1] 

机构地区：[1]齐鲁工业大学财政与金融学院,山东济南250100

出　　处：《计算机仿真》2015年第9期368-371,共4页Computer Simulation

基　　金：山东省重大水利科研与技术推广专项资金资助(SDSLKY201320)

摘　　要：针对海量图像中,可检索的细微特征过多,会使得图像之间的关联性变差,很难根据单一的阀值标准完成图像检索的问题,提出基于细微特征区分的海量小差异图像检索优化模型。在海量图像检索的过程中,提取图像的细微特征,获取图像分割子模块矩阵,统计图像细微特征的分类结果与图像匹配性呈正比例线性关系。对海量图像细微特征进行描述根据细微特征区分结果,实现海量图像检索。实验结果表明,利用改进的海量图像检索模型,能够提高检索效率,满足图像管理的实际需求。For massive images, searchable fine feature is overmuch, so it is difficult to complete the image re- trieval according to the single threshold criteria. This paper puts forward an optimization model for image retrieval based on the fine feature to distinguish the massive small differences. In the process of massive image retrieval, the image fine feature is extracted, the module matrix of subtle image segmentation is obtained, and the massive image retrieval is achieved according to the image fine feature. The experimental results show that the proposed massive im- age retrieval model can improve the retrieval efficiency, meet the practical requirements of image management.

关 键 词：细微特征 图像检索 人工免疫 

分 类 号：F127[经济管理—世界经济]