图书情报工作 ›› 2013, Vol. 57 ›› Issue (03): 112-119.DOI: 10.7536/j.issn.0252-3116.2013.03.021

• 知识组织 • 上一篇    下一篇

采用几何特征的人造物体图像检索研究

李宇翔, 费世英, 李端明   

  1. 西南科技大学经济管理学院
  • 收稿日期:2012-09-17 修回日期:2012-11-18 出版日期:2013-02-05 发布日期:2013-02-05
  • 作者简介:李宇翔,西南科技大学经济管理学院副教授,E-mail:eclili@foxmail.com;费世英,西南科技大学经济管理学院硕士研究生;李端明,西南科技大学经济管理学院教授,研究生导师。

Research on Image Retrieval of Artificial Object Based on Geometric Features

Li Yuxiang, Fei Shiying, Li Duanming   

  1. Institute of Economics and Management, Southwest University of Science and Technology, Mianyang 621010
  • Received:2012-09-17 Revised:2012-11-18 Online:2013-02-05 Published:2013-02-05

摘要:

修正传统图像信息描述方法,在人造物体图像解释过程中采用几何特征和先验知识相结合的方法,提出人造物体图像结构描述以及相应的匹配算法。通过特征点对目标图像的形状轮廓进行离散化和曲线拟合获取离散曲线段的几何特征向量,并根据特征点类型来描述曲线段之间的关系,最终还原人造物体图像结构描述的参数化模型。实验证明,该方法更能够满足人类的视觉特性,而且能够方便计算机还原出图像中的对象形状。

关键词: 图像结构, 几何特征, 图像相似性, 图像检索

Abstract:

To qualify the method of traditional description of image, this paper combines geometric features with prior knowledge in artificial image recognition process, and proposes the method of structure description of artificial object image and the corresponding matching algorithm. It makes the graph structure of object discretization by feature points, and obtains its geometric characteristics with curve fitting. Then it describes the relationships of curves by the types of feature points, and develops the parametric model to restore the object in the image. Experimental results show that the description method has better human vision and can help computer easily restore the shape of the object in the image.

Key words: image structure, geometric feature, image similarity, image retrieval

中图分类号: