基于预测误差直方图的密文域可逆信息隐藏算法  

Reversible Information Hiding Algorithm in Ciphertext Domain Based on Prediction Error and Histogram Translation

作　　者：周杨全 冯柳平 迟灵艺 王郭凯 ZHOU Yangquan;FENG Liuping;CHI Lingyi;WANG Guokai(School of Information Engineering,Beijing Institute of Graphic Communication,Beijing 102600,China)

机构地区：[1]北京印刷学院信息工程学院,北京102600

出　　处：《北京印刷学院学报》2024年第6期1-8,共8页Journal of Beijing Institute of Graphic Communication

基　　金：北京印刷学院学科建设项目(21090323010,21090123009)研究成果。

摘　　要：预测误差扩展是可逆信息隐藏的一种重要方法,预测误差直方图比像素直方图更加集中,峰值点更尖锐。为了增加密文图像的嵌入容量,提出一种基于预测误差直方图的可逆信息隐藏算法,并利用Paillier加密算法的同态特性在密文域中直接嵌入秘密信息。在预处理阶段使用中值预测模型对像素值进行预测,将第一行和第一列作为参考像素,对除参考像素外的所有像素进行预测,生成预测误差直方图,并通过预测误差直方图确定嵌入位置。这种方法充分利用图像像素,增加了目标像素的嵌入位置,从而增加了算法的嵌入容量。测试结果表明,本文的方法有效地提高了嵌入容量。Histogram shifting is an important method for reversible information hiding.The prediction error histogram is more concentrated than the pixel histogram,and the peak point is sharper.In order to increase the embedding capacity,a reversible information hiding algorithm based on prediction error histogram is proposed,and the homomorphic characteristics of Paillier encryption algorithm are used to directly embed secret information in the ciphertext domain.In the preprocessing stage,the median prediction model is used to predict pixel values,using the first row and first column as reference pixels.All pixels are predicted except for the reference pixel to generate a prediction error histogram.The embedding positions are determined through the prediction error histogram.This method makes full use of image pixels to increase the embedding position of target pixels,thereby increasing the embedding capacity of the algorithm.The test results show that the method in this paper effectively improves the embedding capacity.

关 键 词：可逆信息隐藏 预测误差 直方图平移 同态加密 

分 类 号：TP309.7[自动化与计算机技术—计算机系统结构]