基于数字签名方式的图像真伪鉴别算法  

An Image Identification Algorithm Based on Digital Signature Method

作　　者：李晓飞[1,2] 申铉京[1,2] 陈海鹏[1,2] 吕颖达[1,2] 

机构地区：[1]吉林大学计算机科学与技术学院,长春130012 [2]符号计算与知识工程教育部重点实验室(吉林大学),长春130012

出　　处：《计算机研究与发展》2012年第6期1348-1356,共9页Journal of Computer Research and Development

基　　金：国家自然科学基金项目(60973090);吉林省自然科学基金项目(201115025);吉林大学研究生创新基金项目(20111063)

摘　　要：针对图像的真伪问题,提出了一种基于数字签名方式的图像真伪鉴别算法.首先对图像进行预处理来提高图像的抗JPEG压缩性,然后把图像等分成2个区域,分别从2个区域中提取块均值的高5位映射为GF(25)域中的元素来构造RS编码,生成的校验码经Arnold变换置乱后作为数字签名对图像进行真伪鉴别.实验结果表明,该算法不但能够有效地检测出图像内容是否被篡改,并且能够准确定位篡改位置,提高了图像对JPEG压缩的稳健性,有效地区分JPEG压缩操作和蓄意篡改.In view of the question of images＇ authenticity, an effective and robust image identification algorithm is presented based on digital signature. Firstly, image is preproeessed to enhance the capability of images anti-JPEG compression. Then the image is divided into two regions equally, and the top five of image＇s block average value are extracted and mapped to the element of GF （2s） to construct the RS code. The generated check code is scrambled by Arnold transform as digital signature to authenticate the original image. Experimental results show that the algorithm not only can detect whether the content of the image is tampered or not but also can accurately locate the tampered position, which improves the image on the robustness of JPEG compression. Thus this algorithm can effectively distinguish JPEG compression operation from the deliberate tampering.

关 键 词：图像真伪 数字签名 RS编译码 篡改检测 定位 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]