摘要: 提出一种利用显著兴趣点结合颜色矩和距离直方图进行图像检索的方法。该方法将兴趣点作为图像中用户关注的主要视觉线索,包括显著兴趣点检测、距离直方图和颜色矩的特征提取3个步骤,既利用兴趣点的局部颜色特征,又考虑兴趣点间的空间距离关系,克服了传统颜色矩没有空间位置信息的缺陷。实验结果表明,该方法实现简单,能够有效提高图像检索的效率。
关键词:
显著兴趣点,
部颜色特征,
离直方图,
色矩
Abstract: This paper proposes a new image retrieval method which utilizes distinct interest points of the image, and combines color moments and the distance histogram. Interest points of the image are regarded as the main visual cues concerned by users, and it includes detecting distinct interest points, calculating the color distance histogram and color moments. This method not only uses the local color characteristics of an interest point, but also considers space relations of the distance between interest points, and overcomes the traditional color moments’ lacking location information. Experimental results show that the method is simple, and it can effectively increase the efficiency of the image retrieval.
Key words:
distinct interest points,
local color characteristics,
distance histogram,
color moments
中图分类号:
程涛, 侯榆青, 李明俐, 常哲. 一种基于显著兴趣点的图像检索方法[J]. 计算机工程, 2010, 36(18): 171-1763.
CHENG Chao, HOU Yu-Jing, LI Meng-Li, CHANG Zhe. Image Retrieval Method Based on Distinct Interest Points[J]. Computer Engineering, 2010, 36(18): 171-1763.