毕业设计（论文）-基于视觉感知的图像检索精选.doc

摘 要 多媒体技术的快速发展与互联网技术的日益普及，使我们拥有越来越多的数字图像数据。为了能够准确、快速和人性化地从浩瀚的图像数据库中找到用户所需内容，基于内容的图像检索(Content Based Image Retrieval，CBIR)技术应运而生，并已成为国际学术界研究的一个热点。随着基于内容的图像检索技术的渐渐研究深入，一种基于视觉感知的图像检索技术逐渐活跃起来，它是根据人眼的视觉注意机制特点进行检索的，因而可以提高图像检索的精准率。本论文重点围绕基于视觉感知的图像检索进行研究，并提出了一种基于视觉感知的图像检索方法。 该方法使用Itti的视觉注意模型计算得到一个关注度图，在关注度图基础上，使用种子区域增长技术实现对图像中感兴趣物体的自动提取。与一般的图像分割技术不同，种子点的选取和区域增长过程都融合了个体视觉注意程度的信息。最后，本论文通过MATLAB的图形用户界面，构架了一个GUI检索界面，实现了基于视觉感知的图像检索平台。 关键词：基于内容图像检索 高斯金字塔 种子区域增长算法 感兴趣区  ABSTRACT With the development of multimedia technology, the application of Internet and the rapid increment of multimedia database, we have more and more digital images. In order to manage and retrieve those information, the CBIR(Content-Based Image Retrieval) has came into being and emerged to be one of the hot research areas in digital image domain．As the CBIR technology improved，visual perception-based image retrieval technology gradually became activity. It is retrieved based on human visual attention mechanism characteristics. So it can enhance the accuracy of image retrieval .This paper focus on image retrieval based on visual perception .And proposes a method to make it come true. These paper use Itti’s visual attention model to get the attention-degree of the whole image which shows the level of individual attention to every parts of the image. On the base of the concern degree map ,I extract the interested objects in the image automatically with seed region growing technique .And different with the general image segmentation technologies, the process of seed points selection and regional growth combines the information of individual visual attention degree. At the end of the paper, I make a GUI search interface with the MATLAB graphical user interface. And make the search platform based on the visual perception of the image come true. Key words: content-based image retrieval Gaussian pyramid seed region growing algorithm regions of interest 目 录 摘要 1 第一章 绪论 1 1.1 图像检索的应用和技术背景 1 1.2 国内外研究和发展现