摘要
现有图像感知算法侧重于减少相似图像感知哈希值间的差异性,较少关注增大不同图像感知哈希值间的差异性问题。针对这一局限性,提出并实现循环贯序差分和拉普拉斯算子两种基于差分的图像感知哈希算法,并分别在小规模相似图像集合和大规模不同图像集合上进行测试。与现有图像感知哈希算法相比,所提算法体现了图像像素点间的局部关联性,增加了不同图像哈希值间的差异,显著提升了不同图像的区分概率。
The existing image sensing algorithms focus on reducing the difference between similar image perceptual hashing va-lues,and pay less attention to increasing the difference between different image perceptual hashing values.In view of this limitation,two difference based image perceptual hashing algorithms,namely Seq pHash and Laplace pHash,were proposed and implemented,and they were tested on small-scale similar image sets and large-scale different image sets.Comparing with the exis-ting image perceptual hashing algorithms,the proposed algorithm embodies the partial correlation between image pixels,increases the difference between the perceptual hashing values of different images,significantly improves the probability of distinguishing different images.
作者
刘帆
王颖
闫国玉
蒋遂平
LIU Fan;WANG Ying;YAN Guo-yu;JIANG Sui-ping(Institute 706,Second Academy of China Aerospace Science and Industry Corporation,Beijing 100854,China;School of Cyber Engineering,Xidian University,Xi’an 710071,China)
出处
《计算机工程与设计》
北大核心
2021年第3期782-789,共8页
Computer Engineering and Design
关键词
云计算
可搜索加密方案
图像检索
图像匹配
图像感知哈希
cloud computing
searchable encryption scheme
image retrieval
image matching
image perceptual hashing
