摘要
为解决当前图像水印算法难以准确预测相邻两像素之间的差值,使其存在较大的嵌入失真问题,提出基于像素预测误差扩展与优化的极限学习机的图像水印算法。设计像素交叉混淆技术,对水印信息完成置乱,输出水印密文;引入变尺度混沌,对极限学习机进行优化;将混淆后的载体图像像素值作为优化极限学习机的输入信号,对其进行准确预测,引入像素预测误差扩展方法,设计水印信息嵌入机制,将水印数据嵌入到载体图像中;构建水印信息检测机制,利用解密密钥,准确提取出初始水印信息。实验结果表明,与当前水印技术相比,所提算法具有更低的嵌入失真与更高的水印不可感知性。
To solve the problem of large embedding distortion induced by difficultly predicting the difference between the adjacent two pixels in current image watermarking algorithm,an image watermarking algorithm based on pixel prediction error expansion and optimization of the limit learning machine was proposed.The pixel crossing obfuscation technique was designed to scramble the watermark information for outputing watermark cipher.All the pixel values of the carrier image were taken as the input signals of the optimization limit learning machine to accurately predict the pixel value,and watermarking information embedding mechanism was designed by introducing the pixel prediction error expansion method to embed watermark data into the carrier image.The watermark information detection mechanism was constructed to extract the initial watermark information with decryption key.Experimental results show that the propsoed algorithm has lower embedding distortion and higher watermark invisibility compared with the current watermarking technology.
作者
潘强
印鉴
PAN Qiang;YIN Jian(School of Economics and Management, Zhuhai City Polytechnic, Zhuhai 519090, China;School of Date and Computer Science, Sun Yat-Sen University, Guangzhou 510006, China)
出处
《计算机工程与设计》
北大核心
2018年第12期3804-3810,3828,共8页
Computer Engineering and Design
基金
国家自然科学基金项目(61033010
61272065)
广东省自然科学基金项目(S011020001182)
广东省科技计划基金项目(2013B090200006)
关键词
图像水印
像素预测误差扩展
极限学习机
像素交叉混淆
变尺度混沌
回归模型
image watermarking
pixel prediction error expansion
extreme learning machine
pixel cross confusion
mutative scale chaos
regression model