摘要
现有的基于空间域零水印算法缺乏图像内容信息丢失检测的能力,特别是当图像遭受大面积裁切、随机涂抹或图像行列移除等攻击,水印提取算法就会失效。为了提高零水印算法抵抗图像遭受大面积裁切攻击的鲁棒性,提出一种基于四元数分数阶广义拉盖尔图像矩的空间域零水印算法。首先介绍分数阶广义拉盖尔多项式,其次构建四元数分数阶广义拉盖尔矩及不变量,最后设计彩色图像零水印算法。实验结果证明,所提出的零水印算法对常见的几何变换攻击具有一定的鲁棒性,同时对于滤波、有损压缩,特别是对图像大尺度剪切攻击具有较好的检测能力。
The existing spatial domain zero-watermarking algorithms lack the capability to detect loss of image content information,especially when images suffer from large-scale cropping,random scribbling,or row and column removal attacks,it can cause the watermark extraction algorithm to fail.To enhance the robustness of zero-watermarking algorithms against large-scale cropping attacks on images,this paper proposes a spatial domain zero-watermarking algorithm based on quaternion fractional-order generalized Laguerre image moments.Firstly,the paper introduces fractional-order generalized Laguerre polynomials;secondly,it constructs quaternion fractional-order generalized Laguerre moments and invariants based on these polynomials;finally,it designs a color image zero-watermarking algorithm based on the constructed invariants.Experimental results demonstrate that the proposed zero-watermarking algorithm has a certain level of robustness against common geometric transformation attacks and exhibits good detection capabilities for filtering,lossy compression,and particularly for large-scale image cropping attacks.
作者
王晶
贾晓强
WANG Jing;JIA Xiaoqiang(School of Computer Science and Technology,Weinan Normal University,Weinan 714099,China)
出处
《渭南师范学院学报》
2025年第2期87-94,共8页
Journal of Weinan Normal University
基金
陕西省体育局资助项目:智能图像处理及数字版权保护在竞技体育中的应用研究(20240371)
渭南师范学院科学研究计划人才项目:分数阶正交图像矩及其在图像检测中的关键技术研究(2021RC19)
渭南师范学院科学研究计划人才项目:基于超复数分数阶矩的鲁棒性数字图像零水印算法研究(2023RC01)。
关键词
四元数
拉盖尔多项式
零水印
鲁棒性
Quaternion
Laguerre polynomial
zero watermark
Robustness
