摘要
针对DHash或PHash等图像相似检索算法中存在的只依靠一个阈值进行相似判定而造成的不严谨情况,提出了基于双重Hash的图像相似检索算法。算法优先使用耗时较短的DHash进行图像相似判定,采用"阈值+存在误差"的方法进而决定是否需要使用准确度较高的PHash再次进行图像相似判定。实验结果表明,本算法在图像相似检索中的效率与DHash算法相差无几,但较传统PHash或DHash算法能够更准确、完整地检索所有相似图像,具体检索完整性为"本算法> PHash算法> DHash算法"。
Aiming at the imprecise situation caused by the similarity judgment based on only one threshold in the image similarity retrieval algorithm such as DHash or PHash, an image similarity retrieval algorithm based on double Hash is proposed. The algorithm preferentially uses DHash, which takes less time to perform image similarity judgment, and uses the method of “threshold + error” to determine whether it is necessary to use the more accurate PHash to perform image similarity judgment again. The experimental results show that the efficiency of the algorithm in image similarity retrieval is similar to that of DHash algorithm, but it can retrieve all similar images more accurately and completely than the traditional PHash or DHash algorithm. The specific retrieval integrity is "this algorithm > PHash algorithm > DHash algorithm".
作者
尹玉梅
彭艺
祁俊辉
Yin Yumei;Peng Yi;Qi Junhui(Kunming University of Science and Technology, Faculty of Information Engineering and Automation, Kunming 650500, China)
出处
《信息通信技术》
2019年第1期33-38,共6页
Information and communications Technologies
基金
国家地区自然科学基金NO.61761025
