当前位置:网站首页>Reversible watermarking method based on difference expansion
Reversible watermarking method based on difference expansion
2022-07-28 06:16:00 【zzuls】
1. Divide the image into pixel pairs (x, y) x,y∈Z,0≤x,y≤255
Personally think that , This splitting process needs special records , It records which two pixels in the image form this pixel pair , This is to restore the watermark later .
2. Define its integer mean l And the difference h

Formula 1 Formula 2
3. Send secret bit information m It is embedded into the difference by the method of difference expansion h in :
![]()
4. What will be obtained h′ Substitute into formula 2 , Get a new image pixel pair , Form an image after embedding secret information .
5. In order to ensure that the data after reversible hiding does not overflow , Set the carrier image as 8bit Grayscale image of , The conditions for reversible digital watermarking of a pixel pair can be obtained .

Take any pixel pair as an example
The encryption process :
(1) Suppose an image pair is (206,201) namely x=206,y=201
(2)l = floor((x+y)/2) = 203 h=206-201=5
(3) take h Convert to binary 101
(4) Suppose we have a secret message to be embedded in this pixel pair m=1, Then put 1 Add to 101 Of LSB Get on h'
(5) obtain h' = 1011 namely h' = h*2+m = 2*5+1 = 11
(6) hold h' Bring in formula 2 to get :x' = 203+ floor((11+1)/2) = 203+6 = 209 y'=203 - floor((11)/2) = 198
(7) Therefore, the encrypted information is (x',y')=(209,198)
The decryption process :
(1) It is known that (x',y')=(209,198)
(2) Calculation l' = floor((x'+y')/2) = 203 h' = x' - y' = 11
(3)m = LSB(h') = LSB(1011) = 1 // Is o h' The last bit of
(4)h = floor(h'/2) = 5
(5)x = l'+ floor((h+1)/2) = 203 + 3 = 206 y = l'-floor(h/2) = 203-2 = 201
(6) So restore the original pixel pair (x,y) = (206,201) Secret news m = 1
reference :
Jun Tian Reversible Data Embedding Using a Difference Expansion
边栏推荐
- Never leave its origin - bluecms1.6 vulnerability of the controller's shooting range
- Geek challenge 2019-sql injection five questions PW
- The difference between router and switch
- 《On Low-Resolution Face Recognition in the Wild:Comparisons and New Techniques》低分辨率人脸识别论文解读
- Common CTF encryption methods JS
- Deep learning - metaformer is actually what you need for vision
- Alpine, Debian replacement source
- USB Network Native Driver for ESXi更新到支持ESXi7.0 Update 2
- GF(2^8)的加法与乘法计算
- Reinforcement learning - continuous control
猜你喜欢

The difference and relation between TCP and UDP

Notes on the thinking of software analysis and library acquisition of a Taobao customer

基于直方图修改的可逆数字水印方法

深度学习(自监督:CPC v2)——Data-Efficient Image Recognition with Contrastive Predictive Coding

GF(2^8)的加法与乘法计算

深度学习——Pay Attention to MLPs

Tensorflow2.1 basic concepts and common functions

《On Low-Resolution Face Recognition in the Wild:Comparisons and New Techniques》低分辨率人脸识别论文解读

一、语音合成与自回归模型

机器学习之聚类
随机推荐
D2SC-GAN:基于双深浅通道生成对抗网络的课堂场景低分辨率人脸识别
神经网络实现鸢尾花分类
Small program development solves the anxiety of retail industry
Why is the kotlin language not popular now? What's your opinion?
Shutter webivew input evokes camera albums
基于差值扩展的可逆水印方法
使用PowerCli来创建自定义ESXi ISO镜像
The number of password errors during login is too many, and the user is blocked,
ESXi Arm Edition version 1.10更新
强化学习——基础概念
强化学习——多智能体强化学习
C语言EOF的理解
UNL class diagram
无约束低分辨率人脸识别综述三:同质低分辨率人脸识别方法
监控安装ESXi on Arm的树莓派4b的CPU温度
ESXi社区版网卡驱动2022年3月更新
Wechat applet development and production should pay attention to these key aspects
Using neural network to predict the weather
How to improve the efficiency of small program development?
Deep learning (self supervision: Moco V2) -- improved bases with momentum contractual learning