节点文献

数字图书馆图像检索技术研究与实现

【作者】 虞万荣

【导师】 张银福;

【作者基本信息】 国防科学技术大学 , 计算机科学与技术, 2001, 硕士

【摘要】 目前对图像信息的管理技术可以说是基于像素的,而不是基于图像内容理解的。随着图像数据量的剧增,必须提供有效的图像分析和检索机制,使图像管理和检索高效、易行,这使得基于内容的图像研究成为必然。因为仅靠常用的文件标识、关键字或任何与图像相关的文本信息进行索引,局限性太大,无法进行直接基于视觉特性的检索。通过协调好人机分工,让计算机做它擅长的工作,由人来处理非结构化的问题,由此产生了基于内容的图像处理技术。它在不要求理解图像的前提下充分利用其内容的一些可计算特性,诸如颜色、纹理、形状等等,结合其它一些现有的成熟技术,来对图像信息进行存储、管理和检索。 本文首先在Dublin Core的基础上制定了适合我们要求的图像元数据集;详细分析了颜色、纹理、形状等视觉特征的提取和表示方法;探讨了图像视觉特征相似度量的问题,将模糊技术引入直方图的距离度量,分析了几何空间距离度量函数的不足之处,提出了系统中采用的距离函数;针对图像视觉特征向量的多维特性,分析了现有的各种降维技术和多维索引技术。最后综合以上技术,设计实现了一个综合的图像检索系统,集成了元数据检索、全文检索、基于视觉特征检索等多种检索方法,为用户提供友好的使用界面。

【Abstract】 Now the management of image information is still based on pixels,but not on image’s content. With the rapid increase of image data’s amount,we need an efficient analysis and retrieve mechanism for image data to make the management and retrieve of image data efficiently and easily. All these requests make the study of content-based image processing necessary. We let the computer do what it can do,and let ourselves process the non-formalizable problem,we use calculable feature of image’s content(color,texture,shape etc),also with some other mature technology to store,manage and retrieve image information.In this paper,we first established the image metadata used in our system which based on the famous Dublin Core,then we analyzed the abstraction and description visual features of image such as color texture and shape. Next,we discussed the problem of similarity measure of visual feature,imported fuzzy logic into the distance feature and pointed out the disadvantages of geometry space based methods. For multi-dimension vector’s high dimension nature,it’s hard to index with traditional methods,we discussed how to lower the dimension using clustering and KLT transformation. We designed and implemented an image-retrieve system which combined available information retrieving technologies from metadata retrieve,text-retrieve and visual content based image retrieve fields. Also,our system had Web-browser based interface,especially for visual feature,user can visit our site easily and efficiently.

  • 【分类号】TP399
  • 【被引频次】2
  • 【下载频次】284
节点文献中: 

本文链接的文献网络图示:

本文的引文网络