【作者】 张洁婧；

【导师】 刘志镜；

【作者基本信息】 西安电子科技大学 ， 计算机应用技术， 2007， 硕士

【摘要】 随着网络技术、多媒体技术、数据库技术、海量存储技术等技术的发展,数字图像的数量不断增加,使用日益广泛,并成为信息社会中的主要信息资源之一。然而,由于图像往往具有丰富的语义信息和复杂的视觉特征,目前的计算机视觉技术还不能稳定地建立起多媒体对象的语义信息与其视觉特征间的对应关系,使得基于内容的图像检索(CBIR Content-Based Image Retrieval)在检索的准确性上还难以满足实际应用的要求。为了弥补基于内容的图像检索系统性能上的不足,该文对CBIR系统性能优化方法进行了探讨,并通过实验对其优化效果进行验证和评价。本文采用综合特征,并结合人类视觉特点,给出了一种基于颜色-空间信息的检索方法,取得了比一般的单一特征检索方法更好的效果。另外,本文对基于区域的彩色图像检索进行了研究,并给出了一种基于EM和色彩聚类的彩色图像分割方法,使得基于区域的彩色图像检索取得较好的效果。论文最后总结了本文的工作,并提出进一步的研究探索方向。更多还原

【Abstract】 With the development of network, multimedia, database and the mass storage technology, the number of digital images increases and its use wider continual, it become one of the main information resources. Because the images have mass and complex visual features of semantic information, the current computer vision techniques can’t stabile establish the relationship between the semantic information of multimedia objects between its visual features, and the accuracy of the CBIR also difficult to meet the requirements of practical application.In this paper, an integrated features and characteristics of human vision are combined, and a search method based on color-space information retrieval is put forward. The algorithm has the better recognition effect. Otherwise, this paper has a color image retrieval based on a regional study, and gives a color image retrieval algorithm based on region. A method was presented which based on the combination EM algorithm and color clustering, and achieved good results. Finally, the paper summarizes the study and declares the direction of further research and exploration.更多还原