基于邻域LBP算子与块截断编码的图像哈希算法  

Image hashing algorithm based on neighborhood space LBP operator and ordered block truncation coding

作　　者：王彦超[1] WANG Yan-chao(Department of Computer,Pingdingshan College of Education,Pingdingshan 467000,Chin)

机构地区：[1]平顶山教育学院计算机系,河南平顶山467000

出　　处：《计算机工程与设计》2018年第7期2027-2035,共9页Computer Engineering and Design

基　　金：河南省科技计划重点基金项目(102102210416);河南省软科学研究计划基金项目(152400410323)

摘　　要：为提高哈希算法的感知性与鲁棒性,提出一种基于块截断编码与邻域空间LBP算子的鲁棒图像哈希算法。将预处理图像分割为非重叠子块,结合奇异值分解SVD(singular value decomposition),获取二次图像,引入块截断编码机制,输出其高、低电平矩和二进制位图;基于LBP(local binary pattern)算子,设计邻域空间LBP模式,获取位图的特征矩阵;构造量化函数,得到高、低电平矩阵对应的紧凑二值序列,利用主成分分析处理特征矩阵,输出二值序列,组合这些二值序列,获取图像哈希。根据Hamming距离,对图像真伪进行认证。实验数据表明,与当前哈希算法相比,所提哈希算法具有更好的抗碰撞性能与感知鲁棒。To improve the perception and robustness of hash algorithm,a robust image hashing algorithm based on block truncation coding and neighborhood space LBP operator was proposed.The secondary image was obtained by dividing the preprocessing image into non-overlapping sub-blocks and using the singular value decomposition.The high and low level matrixes as well as binary quantized images were got using the ordered block truncation coding mechanism.The neighborhood space local binary pattern was designed based on LBP operator for extracting the feature of the binary quantized images.The compact binary sequences of the high and low level matrices were outputted using the quantization function.The principal component analysis was used to obtain the binary sequences.The image hash was obtained by combining these sequences of feature matrix.The Hamming distance was calculated to finish security authentication of the image information.Experimental data show that the proposed algorithm has better anti-collision performance and perceived robustness compared with the existing hash generation mechanism.

关 键 词：图像哈希 块截断编码 邻域空间LBP算子 奇异值分解 二次图像 用主成分分析 电平矩阵 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]