计算机应用 ›› 2014, Vol. 34 ›› Issue (1): 255-259.DOI: 10.11772/j.issn.1001-9081.2014.01.0255

• 人工智能 • 上一篇    下一篇

基于局部聚类的自适应线性近邻传递分类算法

盛洪波,汪西莉   

  1. 陕西师范大学 计算机科学学院,西安 710062
  • 收稿日期:2013-07-15 修回日期:2013-09-08 出版日期:2014-01-01 发布日期:2014-02-14
  • 通讯作者: 盛洪波
  • 作者简介:盛洪波(1987-),男,湖南益阳人,硕士研究生,主要研究方向:模式识别、图像处理;汪西莉(1969-),女,陕西西安人,教授,博士,主要研究方向:智能信息处理、模式识别、图像处理。
  • 基金资助:

    国家自然科学基金资助项目

Local clustering based adaptive linear neighborhood propagation algorithm for image classification

SHENG Hongbo,WANG Xili   

  1. School of Computer Science, Shaanxi Normal University, Xi'an Shaanxi 710062, China
  • Received:2013-07-15 Revised:2013-09-08 Online:2014-01-01 Published:2014-02-14
  • Contact: SHENG Hongbo
  • Supported by:

    National Natural Science Foundation of China

摘要: 针对线性近邻传递(LNP)分类算法中,由于图像过大时计算复杂度高,以及近邻数目选择不当导致图像分类结果不精确的问题,提出了基于局部聚类的自适应LNP分类算法。该方法对LNP分类算法的改进主要体现在两方面,首先运用quick shift进行局部聚类,得到点簇集,以此点簇集作为建图节点,达到缩小矩阵规模的目的;其次,采用测地距离和欧氏距离之间的关系来动态确定每个点的近邻数。实验结果表明,所提方法在得到较好的分类结果的同时,也极大地缩短了运行时间,提高了效率。

关键词: 半监督, 图模型, 局部聚类, 自适应, 图像分类

Abstract: Linear Neighborhood Propagation (LNP) is a graph-based semi-supervised classification method. It can be used for image classification. LNP has high computational complexity for large images. Furthermore, if the number of neighbors is unsuitable, the classification results will be inaccurate. Therefore, a locally clustering based adaptive classification algorithm for LNP was proposed in this paper. The method improved the LNP classification algorithm in two aspects. Firstly, quick shift was used for local clustering to get the point-clusters. Point-clusters replaced pixels as the graph model nodes. Then the size of matrix was reduced. Secondly, the relationship between the geodesic distance and the Euclidean distance was used to dynamically determine the number of neighbors for each point. The experimental results show that better classification results are obtained by the proposed method and the running time is largely reduced.

Key words: semi-supervised, graph model, local clustering, adaptive selection, image classification

中图分类号: