【摘要】 针对扩频水印抵抗图像剪切能力较差这一缺点,该文将图像转换理论应用到水印领域,提出了一种新的基于幻方变换的算法。在水印嵌入之前,首先将水印图像进行幻方变换,然后对图像进行伪随机扩频得到一个水印序列,最后将水印序列嵌入到载体的低频段部分,实现水印的嵌入。为确保水印信息具有抗剪切能力,在检测阶段,剪切载体的不同部位并且将水印提取出来,计算出水印图像和载体图像之间的相关系数。实验结果表明,当图像受到大面积剪切或者是其他的攻击时,该方法表现出很好的稳健性。更多还原

【Abstract】 Image transformation theory was used with an algorithm based on magic-square transformation then developed to overcome the shortcoming of the poor resisting capacity of spread spectrum watermarking in case of image-cropping.A watermarking image was treated using magic-square transform before embedding watermarking with a watermarking sequence then obtained by spreading the spectrum with a pseudorandom process.The watermarking sequence was embedded into the low frequency band of the carrier to complete watermarking embedding.During detecting,different areas of the carrier were cropped and the watermarking was extracted to ensure the ability of watermarking information to resist image-cropping,with the correlation value between the extracted watermarking image and the carrier image being calculated.Experimental results show that the algorithm can manifest a good robustness when an image is attacked by large area cropping.更多还原