基于本体的Web图像语义标注与检索模型  

Web Image Semantic Annotation and Retrieval Model Based on Ontology

作　　者：张志武[1] 

机构地区：[1]南京邮电大学图书馆,江苏南京210003

出　　处：《情报科学》2015年第4期121-124,131,共5页Information Science

摘　　要：针对传统的基于关键字Web图像检索中的语义缺失问题,结合Web图像的视觉特征和描述文本,利用本体描述Web图像的语义特征,构建了基于本体的Web图像语义检索模型。该模型以领域本体描述Web图像的语义特征,通过自动图像标注技术构建Web图像本体库,用户可以通过输入关键词或者提交示例图像进行图像检索。实验表明,该模型具有较高的图像检索准确率。For remedying the semantics deficiency of the traditional keywords-based Web image retriev- al, combining the visual features and description texts of Web images, using ontology to describe the se- mantic features of Web images, a new Web image semantic retrieval model based on ontology is construct- ed. This model describes semantic features of Web images by domain ontology, and builds a Web image ontology library by automatic image annotation technology. Users can input keywords or submit sample im- ages to retrieval images. The experiments show that this retrieval model has high image retrieval accuracy.

关 键 词：本体 图像检索 自动图像标注 语义特征 

分 类 号：G254.9[文化科学—图书馆学]