节点文献

一种基于区域弦分布的图象检索方法

【作者】 胡恩良

【导师】 林毓材;

【作者基本信息】 云南师范大学 , 基础数学, 2003, 硕士

【摘要】 图象是视觉信息的表现形式和存储载体。随着多媒体、计算机、通信技术及Internet的快速发展,图象的来源越来越来广泛,图象数据库成为组织、表达、存储、查询和利用这些海量图象数据的主要手段。图象数据库研究的一个核心问题是图象检索,其中,基于内容的图象检索技术(CBIR)是当前国内外研究的热点之一。 本文提出了一种基于区域弦分布的图象特征描述,并在此基础上进行了图象检索算法研究。论文的主要工作由图象特征描述和图象检索算法两部分组成。在图象特征描述部分,我们给出了区域弦的定义和区域弦分布的表示,分析了弦分布的性质,并在此基础上进行了图象特征描述。在图象检索算法部分,本文首先作了图象区域划分、区域弦分布的计算;其次对区域作了前景、背景分析并给出了图象间的相似性度量。最后,通过实验和和对实验结果的分析,证明本文研究所提出的方法是有效的。 本文研究成果对基于颜色的图象内容检索具有启发指导作用和实际应用价值。

【Abstract】 Image is the form and the carrier of the visual information.The image source become more extensive along with the rapid development in multimedia technologies , computer science, communication engineering and Internet, in order to organize , describe, store , query and employ these vast image data, image database is one of main tools.The one core problem about image database is the technology of image retrieval, currently, the research of the content-based image retrieval have become one hot spot throughout the world.In this paper,we propose the chord-distribution-based image description ,and research the image retrieval algorithm based on this description.The main work of the paper consists of two part that are the feature description about image and the algorithm of image retrieval.In the first part,we propose the definition of region-chord, the presentation of chord-distribution, and describle image after analyzing the properties of chord-distribution.In the second part,after partitioning image and counting region-chord-distribution,we analyse the forground and background of image to each region and propose the criterion of the similarity of images. Lastly,through experimenting and analyzing the results, it proved that our scheme be effective.To our approach,It be revelatory to color-based image retrieval and be valuable in the application.

【关键词】 图象区域图象检索颜色空间弦分布游程
【Key words】 imageregionimage retrievalchord-distributionrun-length
  • 【分类号】TP391.3
  • 【下载频次】73
节点文献中: 

本文链接的文献网络图示:

本文的引文网络