基于最大意义位(MSB)的图像篡改认证算法

AN MSB-BASED IMAGE JUGGLING AUTHENTICATION ALGORITHM

提出一种用于图像认证的半脆弱水印算法,该算法能够区分出恶意攻击和一般的图像处理。实验表明该算法能准确确定篡改的具体位置,同时对JPEG压缩等常规的图像操作具有较好的鲁棒性。这种算法以16×16个像素点为分块单位对图像进行分块认证,通过小波变换并采用SHA-256产生认证信息,再通过颜色量化的方法,把认证水印信息嵌入到图像的最大意义位,在水印的嵌入过程中,以块为单位对认证信息进行置乱后再进行嵌入,保证了认证的可靠性和水印的鲁棒性,有一定的抗攻击能力。

A semi-fragile watermarking image authentication algorithm is proposed, which is capable of distinguishing hostile attacks and normal image processing. Experiment shows that this algorithm is able to locate accurately the particular position where data is juggled, It also ensures a sound robustness for processing images compressed in common formats such as JPEG. In this algorithm, a block of 16 ×16 pixels is defined as an authentication unit, and authentication information is produced by wavelet transformation and SHA-256. Watermarking informa- tion is embedded into the MSB of an image by color quantization. During the course of embedment, authentication block units are shuffled so as to enhance the reliability of authentication,the robustness of watermarking,and the ability to resist attacking,