1.算法运行效果图预览
灰度图
彩色图
2.算法运行软件版本
matlab2022a
3.算法理论概述
双重水印嵌入算法涉及两个独立的水印:主水印和辅水印,它们可以是灰度图像、二进制序列或其他形式的数据。以下简述嵌入过程:
图像预处理:将彩色图像从RGB空间转换至YCbCr或YIQ等色彩空间,仅对亮度分量(如Y分量)进行水印嵌入,以减少对颜色信息的影响。
图像分割:将亮度分量图像分割成若干个非重叠的M×N大小的子块。
DCT变换:对每个子块进行二维DCT变换,得到DCT系数矩阵。
具体的算法流程图如下所示:
第一重水印:
第二重水印:
4.部分核心程序
%% %基于论文的双重水印算法嵌入 %第1重 %设置嵌入强度 Power = 20; %设置块的大小 Blksize = 8; %改变水印大小,如果太大,则没法全部插入 RR = 32; CC = 32; %第1重 mwk_image1r = func_wk_insert_first(I0r,Marks1r,Power,Blksize,RR,CC); mwk_image1g = func_wk_insert_first(I0g,Marks1g,Power,Blksize,RR,CC); mwk_image1b = func_wk_insert_first(I0b,Marks1b,Power,Blksize,RR,CC); %第2重 mwk_image2r = func_wk_insert_second(mwk_image1r,Marks2r,Power,Blksize,RR,CC); mwk_image2g = func_wk_insert_second(mwk_image1g,Marks2g,Power,Blksize,RR,CC); mwk_image2b = func_wk_insert_second(mwk_image1b,Marks2b,Power,Blksize,RR,CC); %%--------------------------------------------------------------------------------------- %增加攻击 noise0 = 1*randn(size(mwk_image1r)); mwk_image2r = double(mwk_image2r) + noise0; mwk_image2g = double(mwk_image2g) + noise0; mwk_image2b = double(mwk_image2b) + noise0; mwk_image(:,:,1) = mwk_image2r; mwk_image(:,:,2) = mwk_image2g; mwk_image(:,:,3) = mwk_image2b; Msg1r = func_wk_desert_first(mwk_image2r,Blksize,RR,CC,0.8); Msg1g = func_wk_desert_first(mwk_image2g,Blksize,RR,CC,0.8); Msg1b = func_wk_desert_first(mwk_image2b,Blksize,RR,CC,0.8); Msg1(:,:,1) = Msg1r; Msg1(:,:,2) = Msg1g; Msg1(:,:,3) = Msg1b; Msg2r = func_wk_desert_second(mwk_image2r,Blksize,RR,CC,2.2); Msg2g = func_wk_desert_second(mwk_image2g,Blksize,RR,CC,2.2); Msg2b = func_wk_desert_second(mwk_image2b,Blksize,RR,CC,2.2); Msg2(:,:,1) = Msg2r; Msg2(:,:,2) = Msg2g; Msg2(:,:,3) = Msg2b; figure(1) subplot(231); imshow(I0,[]); title('原始图像'); subplot(232); imshow(I1,[]); title('第一重水印'); subplot(233); imshow(I2,[]); title('第二重水印'); subplot(234); imshow(uint8(mwk_image)); title('加入旋转攻击和噪声攻击的图像'); subplot(235); imshow(uint8(Msg1)); title('第一重水印提取'); subplot(236); imshow(uint8(Msg2)); title('第二重水印提取');
标签:彩色图像,RR,CC,wk,水印,matlab,mwk,DCT,Blksize From: https://www.cnblogs.com/matlabworld/p/18168721