基于区域对象的图像检索-计算机应用技术专业论文.docx

摘 要 摘 要 随着数字图像技术的发展，图像检索已经成为一个研究热点，从传统的基于关键字 的图像检索发展到现在广为应用的基于内容的图像检索。基于内容的图像检索主要以图 像的底层视觉内容为特征进行检索，例如：颜色，纹理和形状特征。但是在检索过程中 人们总是将自己的主观感知加入其中，于是就产生了用户的查询需求和底层视觉内容之 间的语义鸿沟。为了减少语义鸿沟，本文提出了一种新的检索方法，它是将图像中用户 感兴趣的区域对象分割出来，然后对这个分割出来的片段进行相似度比较，这样系统就 能很容易地了解到用户的查询需求，减少语义之间的差异。同时我们还将相关系数引入 到颜色特征中，寻求颜色通道之间的相关性，不仅使得检索图像与查询图像保持了颜色 上的一致性，而且还提高了检索精度。 为了不断地学习用户的查询语义，本文将相关反馈引入到图像检索过程中。相关反 馈的过程可以看成是一个图像分类问题，相似的标记为正类，而不相似的标记为负类。 可是，负类图像的个数往往比正类的多很多，这样就存在分类数据的不平衡问题。传统 的两类支持向量机分类器（SVM）就会失效，于是我们分别引入了一类支持向量机 （OCSVM），基于改进AdaBoost的OCSVM集成方法进行实验。通过实验表明，集成方 法比其它方法的性能有了很大的提高，而且使得检索精度得到了提升。 关键词 基于内容的图像检索 区域对象 相关系数 相关反馈 一类支持向量机 I Abstract Abstract With the development of digital image technology, image retrieval has become a hot research topic. From the traditional keyword-based image retrieval to widely used content-based image retrieval. Content-based image retrieval mainly utilizes the low-level visual contents for features to retrieval, for example, color, texture, and shape. However, in the process of retrieval, the users always join their subjective perception in the system, which results in production of a semantic gap between the user’s query and the low-level visual content. In order to reduce the semantic gap, a new retrieval method is proposed in the dissertation. It needs the users to segment out the region they are interested and compare the similarity of those fragments. Hence, the retrieval system can learn the users’ query requirement and reduce the difference between semantics. At the same time, the correlation coefficient is introduced into the color space to find the correlation between the color channels. Therefore, not only the color consistency between the retrieval images and the query image is kept, but also the accuracy precision is also improved. In order to further learn the user’s query semantic, the relevance feedback is introduced into the retrieval process. The process of relevance feedback can be regarded as a classification problem. The relevant images are labeled