文献类型：会议论文

英文题名：Query by concept - An image retrieval method for landscape images

作者：Wang, Yaowei[1]; Su, Limin[2]; Luo, Senlin[1]

第一作者：Wang, Yaowei

通讯作者：Wang, Y.

机构：[1] Department of Electronic Engineering, Beijing Institute of Technology, Beijing, 100080, China; [2] Information School, Beijing Union University, Beijing, 100101, China

第一机构：Department of Electronic Engineering, Beijing Institute of Technology, Beijing, 100080, China

通讯机构：[1]Department of Electronic Engineering, Beijing Institute of Technology, Beijing, 100080, China

会议论文集：5th International Conference on Visual Information Engineering, VIE 2008

会议日期：July 29, 2008 - August 1, 2008

会议地点：Xi'an, China

语种：英文

外文关键词：Query processing - Semantics

摘要：QBE was the most popular manner of CBIR systems. But, "Semantic Gap" is an obstacle to make QBE more efficient. The contribution of this paper was that a framework of QBC (query by concept) was proposed to minimize the effect of "semantic gap". Also, a method was proposed to obtain concepts. An attempt was made with landscape images with "blue sky" and "green grass". Experiment data were about 7000 images collected from Corel database and Internet. The experimental results demonstrated the method was feasible, effective and robust. ?2008 The Institution of Engineering and Technology.