基于颜色和边界方向特征的图像检索

Image Retrieval Based on Color and Edge Direction Features

由于单一特征不足以准确地描述图像特征,提出了一种结合颜色特征和边界方向特征的图像检索方法.针对传统颜色直方图中图像对所有像素具有相同重要性的问题进行了改进,提出了像素加权的改进颜色直方图方法;然后采用非分割图像的边界方向直方图方法提取图像的形状特征,该方法相对分割方法具有简单、有效等特点,并对图像的缩放、旋转以及视角具有不变性.为进一步提高图像检索的质量引入相关反馈机制,动态调整两幅图像相似度中颜色特征和方向特征的权值系数,并给出了相应的权值调整算法.实验结果表明,上述方法明显地优于其它方法.

In this paper, a new method of content-based image retrieval is presented, which integrates color and edge direction feature. First, we research the problems that existed in the conventional color histogram, and present a new color histogram; and then, we extract the shape feature of image by using edge direction histogram of non-segmentation, which is simple and effective, and performs retrieval independent of translation, scaling, and small viewing position variations. At last, similarity of two images is computed by two histogram features. In order to improve retrieval performance, we use a relevance feedback mechanism. This feedback enhances the retrieval effectiveness by dynamically modulating the weights of color-shape information. Experiments show that the proposed method is obviously superior to other methods.