Computer Engineering and Applications ›› 2013, Vol. 49 ›› Issue (4): 166-169.

Previous Articles     Next Articles

Semantic annotation method for images by Web background knowledge

CHEN Shiliang, GUO Xiangdong, DONG Yangyi   

  1. Department of Computer, Mingde College, Northwestern Polytechnical University, Xi’an 710124, China
  • Online:2013-02-15 Published:2013-02-18

结合Web背景知识的图像语义标注

陈世亮,郭向东,董洋溢   

  1. 西北工业大学 明德学院 计算机系,西安 710124

Abstract: In the procedure of semantic annotation methods for images based on content, it is ordinary that identical or similar visual features may have different semantics. Aiming that problem, an image semantic relation model based on Web background knowledge is proposed. The Web page properties related with images are extracted and exploited to compute the semantic relation between images and annotating keywords, and then the semantics of Web images to be annotated are determined. Experiment proves the efficiency of the method proposed.

Key words: image annotation, Web background knowledge, semantic similarity

摘要: 针对基于内容的图像语义标注方法中,相同或相近视觉特征对应语义可能不同的情况,提出了一个结合Web背景知识的图像语义关联模型,利用从Web页面中提取的与图像相关的属性,计算Web图像与标注关键词间的语义相关性,确定待标注Web图像的语义,实验表明该方法具有较好的性能。

关键词: 图像标注, Web背景知识, 语义相似性