基于半脆弱水印的图像内容认证技术  

Image content authentication technology based on semi-fragile watermarking

作　　者：杨珊 李德[1] YANG Shan;LI De(College of Engineering,Yanbian University,Yanji 133002,China)

机构地区：[1]延边大学工学院,吉林延吉133002

出　　处：《延边大学学报（自然科学版）》2020年第1期43-48,共6页Journal of Yanbian University（Natural Science Edition）

基　　金：吉林省教育厅“十三五”科学技术研究项目(吉教科合字[2016]第249号)。

摘　　要：为了认证图像内容的真实性和可靠性,提出了一种用于图像内容认证和篡改检测定位的半脆弱水印算法.首先利用小波变换的系数分量近似生成表征图像内容的特征水印;然后根据图像在小波变换后的能量分布特性,利用伪随机数发生系统确定特征水印的嵌入位置,以此实现水印的自适应嵌入.最后在小波系数中提取出嵌入的水印信息,并与图像生成的水印信息进行比对,以此实现图像的认证功能.实验结果显示,该算法能够区分常规信号的处理操作(如进行高质量的JPEG压缩、添加少量噪声)和恶意篡改操作(如图像剪切),并可对图像的篡改位置进行定位.In order to authenticate the authenticity and reliability of image content,a semi-fragile watermarking algorithm for image content authentication and tamper detection and localization is proposed.Firstly,the wavelet transform is used to approximate the watermarking of image content.Then,according to the energy distribution of the image after wavelet transformation,the pseudo-random number generation system is used to determine the embedding position of the feature watermarking,so as to realize the adaptive embedding of the watermarking.Finally,the embedded watermarking information is extracted from the wavelet coefficients and compared with the watermarking information generated by the image to realize the image authentication function.Experimental results show that the algorithm can distinguish between conventional signal processing operations(such as high quality JPEG compression,adding a small amount of noise)and malicious tampering operations(such as image cutting),and can locate the tampering position of the image.

关 键 词：半脆弱水印 内容认证 离散小波变换(DWT) 自适应嵌入 篡改定位 

分 类 号：TP309.7[自动化与计算机技术—计算机系统结构]