文献类型：期刊文献

英文题名：Research on the image region tagging combined with feature combination and semantic co-occurrence relationship

作者：Wang, Pengfei[1]; Wu, Yanzhang[1]; Liu, Hongzhe[1]; Yuan, Jiazheng[1]; Luo, Yongsheng[1]

第一作者：Wang, Pengfei

通讯作者：Liu, Hongzhe

机构：[1] Beijing Key Laboratory of Information Service Engineering, Beijing Union University, Beijing, China

第一机构：北京联合大学北京市信息服务工程重点实验室

年份：2015

卷号：11

期号：11

起止页码：3957-3965

外文期刊名：Journal of Computational Information Systems

收录：EI(收录号：20153201107638)

语种：英文

外文关键词：Image analysis - Multimedia systems - Search engines - Semantics

摘要：With the development of Internet technology and multimedia technology, images on the network have shown massive growth. The trend of diversification and large quantities of image data have brought serious challenges to traditional image retrieval engines. Regional image annotation concentrates on the relationship between labels and regions, which plays an important role in image retrieval and recognition. This paper shows a method of sparse representation combined with the feature combination and semantic relationship among different parts of an image. The algorithm adds high-level semantic information into the image region annotation process. This reduces the amount of noise involved in the reconstruction process. Tagging accuracy and robustness were improved. Experiments in public image annotation data set have indicated that the algorithm is robust and accurate. ?, 2015, Binary Information Press. All right reserved.