小波纹理与颜色直方图相结合的图像检索技术

【作者】 王莹；

【导师】 彭进业；

【作者基本信息】 西北大学 ， 信号与信息处理， 2010， 硕士

【摘要】 基于内容的图像检索(CBIR)是图像处理领域的一个研究热点。该技术的主要目的是为了客观的查找到用户所要查找的图像。在商业领域和科研领域对基于内容的图像检索都有广泛的研究。其中的关键是图像中的特征提取,特征数据提取的准确与否,直接会影响到图像检索的准确程度。论文主要对小波纹理和颜色直方图相结合的图像检索算法作了深入研究,主要研究成果如下：1.研究了融合Gabor变换和颜色直方图的图像检索算法,分析了此算法在图像检索系统中提取图像的纹理特征和颜色特征的应用,并对此算法进行了MATLAB仿真。2.研究了Curvelet变换提取图像纹理特征的方法,在此基础上,提出了一种融合Curvelet变换和颜色直方图的图像检索算法,并对Curvelet变换和融合Curvelet变换和颜色直方图的算法进行了MATLAB仿真,对融合Curvelet变换和颜色直方图的算法与融合Gabor变换和颜色直方图的算法性能作了分析比较。仿真结果表明,在相同的仿真条件下,融合Curvelet变换和颜色直方图的算法的查准率更高一些。3.对图像检索系统的各个模块进行了设计,使用VC++和MSSQL数据库完成了基于内容的图像检索系统的实现。更多还原

【Abstract】 Content-based image retrieval is a very active research area in digital image processing, whose purpose is to find out the user required image objectively. CBIR is extensively researched in commercial fields and research areas. The key technology of CBIR is to extract the image’s feature data, which has a direct effort on the preciseness of the image retrieval method.The image retrieval algorithm which combine wavelet texture feature and color histogram is deeply studied in this paper. And the main achievements of this thesis are as follows:1. The retrieval algorithm based on Gabor transform and corlor histogram is researched, and analysis the application of this algorithm in CBIR. Then the retrieval algorithm is simulated with MATLAB.2. The applications of the curvelet transform is studied in CBIR to extract texture feature. Based on the curvelet transform, a novel image retrieval algorithm based on curvelet teansform and color histogram is proposed. And the algorithm is simulated with MATLAB. Compared the differences between the proposed method with the method based on Gabor teansform and color histogram, the simulation results show that the new algorithm has better precision.3. Completed a Content-based image retrieval system and designed each modules with VC++ and MSSQL.更多还原