Arnold变换在二值图像置乱应用中若干问题讨论

深圳大学物理科学与技术学院,深圳 518060

计算机应用; 置乱变换; Arnold变换; 二值图像; Arnold逆变换; 图像加密

Discussions on applications of Arnold transformation in binary image scrambling
Liao Rijun, Li Xiongjun, Xu Jianjie, Li Jinlong, Xian Jianbiao, and He Xiaoyu

Liao Rijun, Li Xiongjun, Xu Jianjie, Li Jinlong, Xian Jianbiao, and He XiaoyuCollege of Physics Science and Technology, Shenzhen University, Shenzhen 518060, P.R.China

computer application; scrambling transform; Arnold transformation; binary image; anti-Arnold transformation; image encryption

DOI: 10.3724/SP.J.1249.2015.04428

备注

针对二值图像的特点,讨论了二维Arnold变换在二值图像置乱中的置乱周期问题,提出一种快速Arnold变换策略,即少点置乱法,分析比较Arnold逆变换方法,针对一步恢复原图中直接求矩阵乘方会产生当矩阵元素过大超出运算精度限制时的结果错误问题,提出一种迭代取模伴随矩阵逆变换法.研究结果表明,经Arnold置乱的二值图像,其置乱周期有可能是Arnold变换周期的1/2; 只对二值图像中占像素比例较小的像素进行置乱的变换策略,可提高置乱速度; 对变换矩阵迭代取模后再求伴随矩阵,可得到正确的一步恢复矩阵.二值图像Arnold变换中,综合应用少点置乱法和一步恢复图像方法,可显著提高变换与反变换的速度.

Considering the characteristics of binary images scrambled by Arnold transformation, we discuss the scrambling period for binary images and propose an efficient Arnold transformation strategy, namely scrambling on fewer pixels(SFP). By analyzing and comparing different anti-Arnold transformation methods, we present a one-step image recovery method, the adjoint matrix recovery after iteratively moduloing(AMRAIM), which aims to avoid the error caused by accuracy limitations when directly computing the transform matrix power. Experimental results show that the scrambling period for scrambled binary images may be one half of that of the Arnold transformation depending on the image contents. The calculation speed is improved by scrambling pixels with less proportion in constituting the binary image. By iteratively moduloing the transformation matrix, we obtain the one-step Arnold transformation and its adjoint matrix for anti-Arnold transformation, which recovers the original image in a single step. Combining SFP and ARMAIM in binary image scrambling by Arnold transformation, the computational efficiency is improved.

·