安全且可定位的半易碎图像验证算法  

Secure semi-fragile image authentication algorithm with location capability

作　　者：邵亚非[1] 张利[1] 吴国威[1] 

机构地区：[1]清华大学电子工程系,北京100084

出　　处：《清华大学学报（自然科学版）》2003年第9期1253-1256,共4页Journal of Tsinghua University(Science and Technology)

基　　金：国家自然科学基金资助项目(60172027)

摘　　要：为了有效地解决数字图像内容验证中的定位和安全问题,提出了一种新的基于半易碎水印的验证框架,并给出了DCT域下的实现。通过在图像局部区域提取特征,生成签名并作为水印嵌入,引入纠错编码以提高特征、签名及水印的稳定性,在实现局部定位的同时有效避免了安全上的漏洞,同时在裁剪下仍然保持了验证功能。通过在水印生成中使用有意义的标识和加密手段,使半易碎算法能够抵抗Holliman-Memon的攻击,并在此攻击下仍具有定位功能。The content of digital images is protected with a semifragile authentication framework implemented in the DCT domain. Image features are locally extracted and ECC coded with a signature then generated and embedded as a watermark. Location is achieved without security holes and the authentication still remains active after image cropping. The watermark is combined with a unique block index and cryptography so that modified areas could be correctly located after a HollimanMemon attack. The proposed framework is secure, has low false probability and has both location and anticropping capabilities, so it is superior to other semifragile authentication methods.

关 键 词：图像处理 半易碎水印 特征提取 半易碎图像验证算法 水印嵌入 纠错编码 定位 安全性 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]