现有的基于直方图的可逆水印算法嵌入容量较少,针对这一问题,结合图像放大后像素点的特征,提出一种新的可逆水印算法。利用插值方法将图像放大4倍,既增加了图像的像素个数,又增强了像素之间的相关性;使用上下左右四个像素点计算像素点的预测误差值,并构造直方图,在峰值点处嵌入水印信息;利用相反的过程提取水印,并根据插值原理恢复原始图像。实验结果表明,将图像放大后再嵌入水印,可嵌入水印容量提高了3倍;算法不可见性较好、嵌入容量较大,可用于图像隐藏、信息安全等领域。
The existing reversible watermarking algorithm based on histogram translation has the problem of low embedding capacity. This paper presented a new algorithm combined with the feature of image pixels after amplifying. It was not only increased the number of pixels,but also enhanced the correlation among the pixels by amplifying the picture 4 times with interpolation. It calculated the vertical and horizontal pixels to obtain prediction error value and construct histograms,then embedded watermark information at the peak of histograms. It extracted the watermarking in the reverse process and recovered the original image according to the principle of interpolation. The experimental results show that the embedding capacity has increased by 3 times after being amplified. The presented algorithm shows better invisibility and larger embedding capacity,and also can be applied to the domain of image hiding and information security.