基于兴趣点局部分布特征的图像检索方法

A Method for Image Retrieval Technique Based on Local Distribution Features of Interest Points

提出了一种基于兴趣点颜色和空间分布特征的图像检索方法。该方法把图像内容看作为由若干兴趣点组成的集合,首先利用小波系数的空间方向树特性来检测兴趣点,然后利用基于兴趣点的环形颜色直方图和空间离散度来描述图像的特征,最后用加权特征距离来估计图像内容的相似度。同时,通过利用环形颜色直方图和空间离散度作为图像特征保证了该算法能够对图像的尺度变化、旋转变化和平移变化具有很好的抑制能力。在含有1000幅图像的数据库上所做的一系列实验表明,该算法与其它基于兴趣点的方法相比,能够更准确和高效地查找出用户所需的图像,明显地提高了检索精度。

A novel method for image retrieval based on color and spatial distribution features of interest points was presented. The content of one image is looked as an aggregation of some interest points. Firstly,we present a wavelet-based detector, which uses the space-tree property of the transform coefficients to estimate the interest points. Secondly, to provide geometric invariant image matching,annular color histogram and spatial cohesion based on interest points are presented to describe image features. Finally,the weighted feature distance is used to discriminate the similarity between two images. The annular color histogram and spatial cohesion are used to make this method remain invariant to the image＇s scale,rotation and translation. A series of experiments based on an image database consisting of 1000 images are performed to confirm the effectiveness of our method. In these experiments, our method provides more accurate and efficient retrieval performance comparing with other interest points-based retrieval method.