基于缩略图和EMD的无分割空间彩色图像检索方法

辽宁石油化工大学学报 ›› 2009, Vol. 29 ›› Issue (1): 73-75.

基于缩略图和EMD的无分割空间彩色图像检索方法

李丽君

辽宁石油化工大学理学院,辽宁抚顺113001

收稿日期:2008-10-16 出版日期:2009-03-25 发布日期:2017-07-05
作者简介:李丽君(1979-),女,辽宁沈阳市,讲师,硕士

Spatial Color Image Retrieval Without Segmentation Based on Thumbnails and the Earth Mover's Distance

LI Li－jun

chool of Sciences, Liaoning University of Petroleum & Chemical Technology, Fushun Liaoning 113001, P.R.China

Received:2008-10-16 Published:2009-03-25 Online:2017-07-05

摘要/Abstract

摘要： 提出了一种没有引入任何分割过程的空间彩色图像检索方法。首先结合人眼视觉感知特性,在
CIELab颜色空间提取查询图像的缩略图,然后使用EMD 方法进行图像缩略图间的匹配。在匹配过程中基于a
contrario方法提出了一种无监督的匹配原则。实验结果表明,此算法不仅检索过程简单,鲁棒性强,而且检索效率明
显高于基于区域分割的图像检索方法。

关键词: 缩略图 , EMD , 背景模型 , acontrario方法

Abstract: A spatial color image retrieval method without any segmentation step was proposed. Thumbnail of query image was extracted in CIELab color space considering the human visual characteristic.Thumbnails of images were matched using EMD.An unsupervised matching criterion was presented based on a contrario approach. The experimental results show that our image retrieval is more simple and robust, and has better performance than image retrieval based on regions segmentation.

Key words: Thumbnail , Earth measure's distance , Background model , A contrario approach

李丽君. 基于缩略图和EMD的无分割空间彩色图像检索方法[J]. 辽宁石油化工大学学报, 2009, 29(1): 73-75.

LI Li－jun. Spatial Color Image Retrieval Without Segmentation Based on Thumbnails and the Earth Mover's Distance[J]. Journal of Liaoning Petrochemical University, 2009, 29(1): 73-75.

[1]	付春,孙祥磊,王昆,等. 基于FastICA改进EMD的算法研究[J]. 辽宁石油化工大学学报, 2017, 37(5): 67-70.
[2]	侯操,孙小平. 列车视频监控中视频遮挡报警算法的研究与应用[J]. 辽宁石油化工大学学报, 2015, 35(3): 61-64.