基于图像内容认证的Contourlet域半脆弱水印算法

Contourlet Image Content Authentication Based on Semi-fragile Watermarking Algorithm Domain

本文针对小波变换不能有效表达二维信号的状况,提出了一种基于图像内容认证的Contourlet域半脆弱水印算法,算法修改了小波域的HVS模型,将修改后的模型应用于Contourlet变换域,提取图像的颜色、边缘、纹理特征作为水印信息.水印嵌入时选择变换后能量最大的方向子带,将该方向子带利用混沌映射置乱后嵌入水印.实验证明本算法提高了载体图像的透明性与水印的安全性.认证检测时,无需原始图像和原始水印,实现了盲认证.算法能够抵抗JPEG、JPEG2000压缩等一般的图像处理操作,并对恶意篡改报警,定位篡改部位.滑动窗口滤波消除了一般图像处理操作产生的虚假的报警点.

According to wavelet transform non effective expression two-dimensional, we proposed an algorithm based on the content of images in the contourlet domain, modified the HVS model in the wavelet domain to adapt contourlet transform and extracted the watermark. Watermark is embedded in the largest energy sub-band after eontoudet transform,increased transparency and security of watermark. Image authentication and watermark extract needed no information about the original watermark and original image, achieved a blind certification. Algorithm can resist the JPEG, JPEG2000 compression and other common image processing operations, give an alarm for malicions tampering, and locate tampered parts. Sliding window filter eliminates the general image processing operations generated a false alarm.