基于IPVO的高效图像可逆认证方法  

Efficient image reversible authentication method based on IPVO

作　　者：张汉华 黄方军[1,2] ZHANG Hanhua;HUANG Fangjun(School of Computer Science and Engineering,Sun Yat-sen University,Guangzhou 510006,China;School of Cyber Science and Technology,Sun Yat-sen University,Shenzhen 518107,China)

机构地区：[1]中山大学计算机学院,广东广州510006 [2]中山大学网络空间安全学院,广东深圳518107

出　　处：《中山大学学报（自然科学版）（中英文）》2024年第5期106-114,共9页Acta Scientiarum Naturalium Universitatis Sunyatseni

基　　金：国家自然科学基金(U2336208,62072481)。

摘　　要：提出了一种基于小波域预测误差直方图平移的图像可逆认证方法。首先,对待认证图像进行一次哈尔小波变换;其次,将变换得到的低频子带和高频子带分别分块,并根据块内的系数值和块的位置生成认证码;然后,使用改进的像素值排序(IPVO)方法将认证码嵌入到分块中,通过哈尔小波逆变换得到包含认证信息的图像;在认证阶段,通过对比分块提取认证码和生成认证码来实现篡改检测功能,并可无损地恢复原始图像。实验结果表明,该方法提供了更准确的检测结果,并确保了一定水平的图像质量。A reversible image authentication method based on wavelet-domain prediction error histogram translation is proposed.First,the image to be authenticated performs a Haar wavelet transform.Second,the low-frequency and high-frequency subbands obtained from the transform are separately partitioned into blocks,and an authentication code is generated based on the coefficient values and locations of these blocks.Then,the authentication code is embedded into the blocks using the improved pixel value ordering(IPVO)method,and the image containing the authentication information is reconstructed using the Haar wavelet inverse transform.In the authentication stage,the tampering detection function is realized by extracting the authentication code and generating the authentication code.If the image is not tampered with,the original image can be recovered losslessly.Experimental results show that the method provides more accurate detection results and ensures a certain level of image quality.

关 键 词：图像 可逆认证 脆弱水印 篡改检测 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]