基于DCT变换的自适应图像水印实现

Adaptive image watermark realization based on DCT transform

提出了一种基于DCT变换的自适应水印嵌入和检测方法。这种方法可以根据分块图像矩阵的特征,利用DCT变换的特点和自适应算法,在不同的分块区域中嵌入不同强度的水印,在图像的每个分块区域内都能实现水印强度和透明性的自适应调节。在检测水印时,还采用了一种有四级优先级的图像恢复方法,可以更好地恢复水印图像。利用这种方法成功地进行了水印的嵌入与检测,并对水印图像进行了量化攻击、JPEG压缩攻击、高斯噪声攻击和裁剪攻击检验。实验表明,这种方法可以在保证透明性的条件下,具备很好的鲁棒性,有较好的实用价值。

A watermark embedding and detecting adaptive measure based on discrete cosine transform （DCT） was introduced. Different intense watermarks were imbedded to different ranges on adaptive algorithm based on the character of segmental image matrix and the property of DCT. Adaptive regulation of intensity and transparence was realized in every segmental area. An image restoration method with four priority levels was introduced for better detecting effect. Imbedding and detecting experiments of watermark were successfully realized with this method. Quantization attack, JPEG compression attack, gaussian noise attack, cutting attack were respectively added to the experiment. The experiments showed that the method could ensure both the quality of watermark transparence and a nice robustness. Therefore, it enjoys a good practicability.