基于奇异值分解的半易损水印算法

随着数字图像在报刊杂志、医院、法庭中的广泛应用，越来越需要一种有效的图像认证方法，数字水印技术为上述问题提供了一个潜在的解决方案。本文提出了一种基于分组奇异值分解（SVD）的半易损水印技术，算法将经过伪随机排序的二值图像通过量化策略嵌入到分组SVD分解中最大的奇异值点，提取水印信号无需使用原始图像。仿真实验表明水印是不可察觉的，可将JPEG有损压缩同恶意攻击区分开来，能够准确地定位被篡改的图像内容。

SVD Based Semi-fragile Watermarking Method

With the application of digital images in deferent area., it is becoming easier to modify or forge image information using widely available editing software. Digital watermarking techniques have been recognized as an effective solution to image authentication. A semi-fragile watermarking technique based on block singular value decomposition(SVD)is proposed. Binary watermark image of pseudo-random permutation is embedded into the biggest singular value by quantization process. watermarks can be extracted without the original pattern. Experimental results show that the proposed scheme can be used to prevent malicious attacks but permit JPEG format compression and to locate alteration made on the image.