基于DCT域的自适应图像水印算法

An Adaptive Image Watermarking Algorithm Based on DCT

提出了一种DCT域自适应图像水印算法。嵌入水印的过程中不断地搜索合适的强度因子,根据JPEG亮度量化表来确定中频系数嵌入强度的比例关系,并引入了一个优于PSNR和MSE的方法来评价含水印图像失真。若图像质量不满足所期望接近的失真度,用二分法不断地调整强度因子的值,以达到水印的最优嵌入,从而水印图像信息分别以不同的强度嵌入到各中频系数中。实验结果表明该水印算法对常见的信号处理具有较好的稳健性。

An adaptive image watermarking algorithm based on DCT is proposed here. The embedding process is modified continuously in order to find appropriate scaling parameters. According to the JPEG brightness table, the proportion of embedding strength of mid-frequency coefficients is confirmed. And a new universal image quality index which is superior to PSNR and MSE is introduced here to analyze the embedded image distortion. If it can＇t be close to the expeeted distortion, the scaling parameter is adjusted by diehotomy so as to obtain optimal embedding effect. So watermark are embedded into the mid-frequency coefficients with different strength. Experimental results show that the adaptive algorithm is invisible and robust to certain common distortions.