文本检索方法驱动的海量图像数据库目标匹配研究

On the Object Retrieval of Large-Scale Images Based on Text Retrieval

针对当前基于文本检索方法的图像目标对象匹配技术无法适应海量图像数据库检索的问题,本文提出一种有效可行的海量图像数据库的检索方法,并给出了该系统的构建框架。用户通过在图像中选择一块区域作为检索的目标对象提交给系统,它将从图像数据库中检索出包含有相同或相似目标对象的图像,将其排序后返回给用户。实验表明,本文提出的方法具有检索准确率高、响应时间短等特点,是一种有效的海量图像数据库检索方法。

In this paper, we present a novel approach to tackle a large-scale object retrieval problem, and demonstrate an object retrieval system architecture based on this method. The user supplies a query object by selecting some region of a query image, and the system returns a ranked list of images retrieved from a large corpus,which contains the same object. The experiments show that our approach outperforms other models in efficiency and accuracy.