基于分层聚类相关反馈算法的图像检索技术研究

On the image retrieval method based on hierarchical clustering relevance feedback

基于内容的图像检索研究的目的是实现自动地、智能地检索图像,研究的对象是使查询者可以方便、快速、准确地从图像数据库中查找特定图像的方法和技术。通过把分层聚类策略与传统的相关反馈算法相结合,提出一种新的图像检索方式,并通过实验加以验证。

The present paper aims to： （1） integrate hierarchical clustering with the traditional relevance feedback to derive a new image retrieval method, and （2） test its capability and feasibility, and its advantages over other methods. The new method is tested to be capable of retrieving a certain image with more convenience, greater rapidity and higher accuracy, and thus excels in its higher efficiency in computing, and higher precision ratio, accuracy and efficiency in retrieving.