一种新的基于特征的图像内容认证方法

A New Feature-Based Image Content Authentication Algorithm

该文提出了一种新的基于特征的图像内容认证方法 .它首先利用小波变换多尺度边缘检测方法抽取出光滑分量和边缘特征作为生成签名的图像特征集 .这可以保证特征集与图像之间的一一映射关系 ,避免伪造攻击 .然后它将光滑分量量化后生成 12 8位的摘要 ,并将边缘特征以水印方式嵌入原图 .验证图像时需同时通过与摘要的对比以及与水印特征的对比来确定内容是否被篡改 .实验表明 ,该算法可有效地检测出恶意篡改及其发生的位置 ,并可容忍由压缩、滤波等操作引起的失真 .即使图像受一定噪声污染 ,它也可通过验证 .

This paper presents a new feature based image content authentication algorithm. First, it draws the smooth component and edge character from the image with the algorithm of dyadic wavelet transform multi scale edge detection, and makes them the image feature set that produces the digital signatures. That can ensure the one to one mapping relationship between the feature set and the image, and avoid the counterfeit attack. Then it quantizes the smooth component to create 128 bits digest, and embeds the edge character in the original image as a watermark. Image authentication needs to contrast with the digest and watermark character to confirm whether the content is tampered. The experiment proves that this authentication algorithm can effectively detect the event and the location when the original image content is tampered viciously, and can tolerate the damage produced by compression, filter and some other image processing. Even if the image is degraded by some noise, it can still pass through the authentication.