利用灰色系统理论确定ICA解密图像的顺序  

Application of Grey System Theory to Determining Order of ICA Decrypted Images

作　　者：谢松云[1] 王颖[1] 谢玉斌[2] 李海波[1] 

机构地区：[1]西北工业大学电子信息学院,陕西西安710072 [2]中国工程物理研究院电子工程研究所,四川绵阳621900

出　　处：《西北工业大学学报》2015年第1期153-158,共6页Journal of Northwestern Polytechnical University

基　　金：国家自然科学基金(61273250);西北工业大学种子基金(Z2013085)资助

摘　　要：针对基于ICA的图像加密解密算法中,解密输出图像顺序不确定问题,提出一种利用灰色系统理论进行图像特征匹配的方法。利用主成分分析提取图像的主分量数据,建立灰预测模型GM(1,1),提取模型的灰参数作为匹配特征;将解密图像的灰参数作为比较序列,对明文图像的灰参数作灰关联分析,关联度最大的图像对就是匹配的明文图像与解密图像。分别选取独立性强和关联性强的明文-解密图像集各1组,进行图像匹配实验,获得了准确的匹配结果。该方法对图像具有普适性,计算复杂度小,确定ICA解密图像顺序快速、准确。Aiming at removing the permutation ambiguity of decrypted image sequence outputs resulting from image encryption and decryption algorithm based on Independent Component Analysis （ICA）, we propose an image fea- ture matching method using grey system theory. First the principal components are extracted by using Principal Component Analysis （PCA）, and the grey parameters are obtained as features by establishing grey prediction mod- els GM （ 1,1 ）. Then take the grey parameters of decrypted images as comparative sequences and apply grey rela- tional analysis to those of original images; the image having the greatest correlation with the decrypted image is the matched one. Matching experiments are performed on selected original-decrypted images with obvious independence or strong correlation, and results with precision are obtained. The proposed method is of universality as well as low computational complexity, and can determine the permutation of ICA decrypted images rapidly and accurately.

关 键 词：独立成分分析 解密图像顺序 主成分分析 灰预测模型 灰关联度 

分 类 号：TN911.73[电子电信—通信与信息系统]