该目录包含了基于可移植的C语言的数字水印算法的代码。这些代码具有示范

性的意义，且有一定的鲁棒性。



需要安装NetPGM的程序包（用来存取pgm格式图像），然后才能在linux下面编译。

NetPGM是图像文件的输入/输出软件包。可以在网上搜索得到。我们已经对

256灰度级，512×512大小的图像进行了测试。程序编译后的调用格式如下（以encode为例）：

wm_cox_e -s cox.sig -o wm_image.pgm image.pgm

其中，-s表示要添加的水印信号，cox.sig是信号文件。

-o表示输出图像文件名为wm_image.pgm，

image.pgm为输入文件名。



几个辅助程序：

cmp_pgm 计算差分图像，PSNR等

cmp_dct 计算全帧DCT域的差分图像

cmp_dct8x8 计算8×8大小的DCT域差分图像

比如说，为了产生两个PGM图像的差分图像并计算它们的PSNR，命令如下：

cmp_pgm -p -i original.pgm -o diff.pgm watermarked.pgm



产生水印信号：

gen_cox_sig -o cox.sig

程序将产生适合512×512大小，8-bit灰度的pgm图像的水印信号。



水印提取：

wm_cox_d -s cox.sig -i lena.pgm -o cox.wm cox_lena.pgm

请注意，cox的算法需要原始图像做参考来提取水印，压缩后的水印将存储在cox.wm中。-This directory contains the C-based portable digital watermarking algorithm code. The code has a demonstration

Of meaning, and there is a certain degree of robustness.



NetPGM need to install the package (to access the pgm format images), and then to compile in linux below.

NetPGM is the image file input/output package. Can search by. We have already

256 gray level, 512 × 512 size images were tested.

Procedure call format, compiled as follows (in encode an example):

wm_cox_e-s cox.sig-o wm_image.pgm image.pgm

Which,-s that you want to add the watermark signal, cox.sig is the signal file.

　　 -O said output image file name wm_image.pgm,

　　　image.pgm for the input file name.



Several auxiliary programs:

cmp_pgm calculate the difference image, PSNR, etc.

cmp_dct full frame DCT domain calculation of difference image

cmp_dct8x8 calculate the size of 8 × 8 DCT domain differential image

For example, in order to produce two PGM images and calculate their differ