摘要： 重复图像检测是自动图像标注中经常遇到的问题之一。该文在讨论大规模图像数据库的基础上，提出一种基于Hash值的重复图像检测算法。该算法不依赖于具体图像特征，通过建立索引能快速寻找到重复图像，有效提高了查准率。实验结果表明，该算法是可行的，可以应用到其他各种场景。

关键词: 图像检测, 重复图像, 图像内容, Hash值

Abstract: Duplicate image detection is one problem in automatic image tagging. On the basis of discussing the large-scale images database, a duplicate image detection algorithm is proposed, which does not depend on the specific characteristic of images. This algorithm can find the duplicate image by setting up the index, and promote the accuracy of query effectively. Experimental results show the algorithm is feasible and can be used into other relevant cases.

Key words: image detection, duplicate image, image content, Hash value

中图分类号: 

• TP391