摘要
分析了CBIR系统的结构模式,探讨了基于形状特征的图像检索系统构建问题。重点研究了形状描述、特征向量索引以及特征相似性度量及匹配等相关技术,应用轮廓点与兴趣点之间的空间分布关系构造形状描述函数并提取图像特征。实验结果表明,系统在动物形状测试集中具有较高的检索效率。
In this paper, the structure pattern of CBIR was analyzed, and the construction of shape - based retrieval sys- tem was discussed. The research mainly focused on the technologies of shape description, feature vector index and similarity mea- sure and match. The contour and feature points were used in constructing the description function and extract the shape features. Retrieval results in the animal shape dataset showed the effectiveness of the system.
出处
《现代情报》
CSSCI
2014年第5期79-82,87,共5页
Journal of Modern Information
基金
国家社科基金重大资助项目"图书
博物
档案数字化服务融合研究"(项目编号:10&ZD134)
关键词
数字图书馆
基于内容的图像检索
形状分析
系统构建
digital library
content- based image retrieval
shape analysis
system construction