基于内容图像检索中的特征性能评价

Feature Performance Evaluation in Content-Based Image Retrieval

在基于内容的图像检索中,不同图像特征反映了图像各个侧面的内在特性,因此,在使用图像特征进行检索时存在多种相似性度量方法.特征以及特征间相似性度量方法的选取是当前CBIR研究的一个重要课题,评估了CBIR系统中使用的图像特征在不同相似性度量方法下及多种特征在不同图像库上的检索性能,为CBIR系统的设计和实现提供一定的依据.通过实验发现,图像特征的检索性能不仅同相似性度量方法有关系,同时与图像库也有密切的关系.

In content based image retrieval(CBIR), image has various inherent aspects, which reflect its contents, therefore, it needs different similarity measures to improve the retrieval performance. Features and their corresponding similarity measure selections are important research topics in CBIR. In this paper, image features used in CBIR system on different similarity measures are evaluated, and a variety of features on different image collections are also evaluated, which can be the basis of the CBIR system development. According to the experimental results, feature's retrieval performance is not only related to similarity measure but also relies on image collection.