基于特征点组合聚类的图像检索新方法

由于特征点能对图像局部特征进行合理描述,有效使用特征点实现基于内容的图像检索成为当前计算机视觉领域中的热点问题.针对该问题,提出一种基于特征点组合聚类的图像检索新方法.该方法包括特征点组合聚类算法,以及基于该算法的局部颜色直方图构建策略.与现有的基于特征点和局部颜色直方图的检索方法相比,该方法能有效解决当前方法对特征点位置信息及特征点中心过度依赖的问题.从公共图像库上的实验结果可以看出,该方法与现有方法相比具有较高的检索精度.

A novel image retrieval method based on combinatorial clustering of feature points

Because feature points can represent the local feature of images effectively, how to realize content-based image retrieval by feature points is a hot issue in the computer vision. So, a novel image retrieval method based on the combinatorial clustering of feature points is proposed. The proposed method consists of combinatorial clustering algorithm of feature points and local color histogram building approach. Compared with existing methods based on feature points and local color histogram, the proposed method can solve the problem that existing methods are sensitive to the position and the center of feature points. According to experimental results on public image database, it can be seen that the proposed method is more effective than existing methods.