基于均匀量化的二维多尺度排列熵算法
Two-dimensional multi-scale permutation entropy algorithm based on uniform quantization
王大铭 1史鹏飞 1雷一航 1边皓冉 1梁敏 1常利伟1
作者信息
- 1. 山西财经大学信息学院,山西 太原 030006
- 折叠
摘要
为了解决将排列熵算法扩展到二维时,子序列中相等值会导致某些排列模式的概率增加的问题,提出了一种基于均匀量化的二维多尺度排列熵(MUPE2D)算法.算法通过基于均匀量化重新定义排列模式,消除了相等值对计算的影响.使用MUPE2D算法对各种合成纹理、MIX2D(p)图像和加密图像进行了研究,结果表明,即使图像中存在大量等值,MUPE2D算法也能有效量化加密图像的复杂性和信息隐藏能力.综上所述,MUPE2D算法为评估图像复杂度提供了一种有效的手段.
Abstract
In order to solve the problem that equal values in the subsequence lead to an increase in the probability of cer-tain ordinal patterns when extending the permutation entropy algorithm to two-dimensional domain,a two-dimensional multi-scale permutation entropy based on uniform quantization(MUPE2D)algorithm based on uniform quantization was proposed.By redefining the ordinal patterns based on uniform quantization,the effect of equal values on the calculation was eliminated.The MUPE2D algorithm was used to investigate various synthetic textures,MIX2D(p)images and en-crypted images.The results reveal that MUPE2D is able to effectively quantify the complexity and information conceal-ment capabilities of encrypted images even if there are a large number of equal values.Consequently,the MUPE2D algo-rithm provides an effective means for evaluating images.
关键词
二维排列熵/多尺度熵/加密图像分析/纹理分析Key words
two-dimensional permutation entropy/multi-scale entropy/encrypted image analysis/texture analysis引用本文复制引用
出版年
2024
国家科技期刊平台
NSTL
万方数据