[image hiding] digital image watermarking method technology based on hybrid dwt-hd-svd with matlab code

1 Content introduction

This paper presents a method based on discrete wavelet transform (DWT),Hessenberg decompose (HD) And singular value decomposition (SVD) Image watermarking method . In the process of embedding , Multilevel processing of the original carrier image DWT decompose , The subband coefficient obtained is taken as HD The input of . While creating the watermark SVD To operate , The watermark is embedded into the main image by scaling factor .. Under various deception attacks , Compare the proposed method with other methods , Experimental results show that , This method has good robustness and invisibility to watermark .

The robustness and invisibility of image are two important factors to evaluate the effectiveness of watermarking technology   Main indicators . Watermarking technology can be roughly divided into three categories , That is, robust watermark 、 fragile   Watermarks and semi fragile watermarks  [1]. Robust watermarking is very important for the protection of image data ,  Because it will not significantly reduce the visual quality of the watermark image , And it can resist   All kinds of attacks . therefore , Robust watermarking is widely used in copyright protection and ownership verification .  Fragile watermarking is only used to ensure the integrity of the image , It can't verify all the actual   power . Semi fragile watermarking combines the advantages of fragile watermarking and robust watermarking , To check   Detect unauthorized operations , While maintaining robustness to authorization operations . Besides ,  The embedding process can also be completed in the transformed domain , Such as singular value decomposition （SVD）、  Discrete cosine transform （DCT）、 Discrete Fourier transform （DFT）、 Discrete small   Wave transformation （DWT）. Studies have shown that , Human vision is more sensitive to medium and low frequency coefficients . therefore , change   The operation method in the commutation field has good performance , Especially when the watermark is embedded in low   Within the frequency range . be based on  DWT  The watermark method has high resolution 、 Energy pressure   Good contractibility 、 High visual quality , It can be used for image watermarking . However , be based on  DWT  The watermark is difficult to resist geometric attacks . This drawback can be divided by matrix   The solution method extracts the geometric features of the image to solve , be based on  DWT  And matrix decomposition   This method is widely used in image watermarking , So that the watermark can withstand a certain range   Degree image processing and image geometric attack  [2]. The most commonly used matrix decomposition in watermarking   Methods include  SVD  and  Hessenberg  decompose （HD）. Besides , Robustness and   Invisibility is the two main indicators to evaluate image watermarking , Performance between the two   Balance is very challenging . This paper makes use of  FOA  Algorithm to Algorithm   The parameter factor is improved , A trade-off is made between robustness and invisibility .  This paper proposes a combination  DWT、HD  and  SVD  Image Watermarking Algorithm  [3].  Performance test results show that , This method has high robustness and invisibility ,  And there is no constraint on the size of the watermark .

The input image is decomposed by multi-level wavelet , Generate three high frequency bands  HH、 LH、HL  And a low frequency band  LL, Pictured  1  Shown . The low frequency band is   Column is the wavelet decomposition series corresponding to the minimum resolution and maximum scale of the original image   The best approximation of image . Its statistical characteristics are similar to the original image , Most of the image can   The quantity is concentrated here . High frequency band is the content of images at different resolutions and scales   Details of . Lower resolution , The higher the proportion of useful information . chart   Like wavelet decomposition is divided into several levels , For peer images , Low frequency sub image  LLj  most   important , The second is  LHj  and  HLj, And high frequency sub image  HHj  Relatively unimportant .  For different levels , The higher level is important , It doesn't matter who is lower , Pictured  1  Shown .  Wavelet theory makes further decomposition possible , Until the size of the subband reaches   Watermark standard . Compared with other subbands ,LL  Better attack performance , for example   Filter 、 Compression attack . This characteristic makes  LL  Subband becomes the most important part of robust watermarking   Good candidate  [4].

2 Simulation code

<span style="color:#333333"><span style="background-color:rgba(0, 0, 0, 0.03)"><code>%% This function applies attacks <span style="color:#ca7d37">on</span> images</code><code>%  Input: image, attack type, attack <span style="color:#dd1144">parameters</span> (different <span style="color:#ca7d37">for</span> each attack)</code><code>function [watermarked_image] = Attacks(watermarked_image,attack,param)</code><code><span style="color:#ca7d37">switch</span> attack</code><code>    <span style="color:#ca7d37">case</span> <span style="color:#dd1144">'No Attack'</span></code><code>    <span style="color:#ca7d37">case</span> <span style="color:#dd1144">'Median'</span></code><code>        watermarked_image = medianAttack(watermarked_image,param);</code><code>    <span style="color:#ca7d37">case</span> <span style="color:#dd1144">'Gaussian noise'</span></code><code>        watermarked_image = noiseGauss(watermarked_image,param);</code><code>    <span style="color:#ca7d37">case</span> <span style="color:#dd1144">'Salt and pepper noise'</span></code><code>        watermarked_image = noiseSaltPepper(watermarked_image);</code><code>    <span style="color:#ca7d37">case</span> <span style="color:#dd1144">'Speckle noise'</span></code><code>        watermarked_image = noiseSpeckle(watermarked_image);</code><code>    <span style="color:#ca7d37">case</span> <span style="color:#dd1144">'Sharpening attack'</span></code><code>        watermarked_image = sharpenAttack(watermarked_image,param);</code><code>    <span style="color:#ca7d37">case</span> <span style="color:#dd1144">'Rotating attack'</span></code><code>        watermarked_image = rotatAttack(watermarked_image);</code><code>    <span style="color:#ca7d37">case</span> <span style="color:#dd1144">'Motion blur'</span></code><code>        watermarked_image = motionAttack(watermarked_image);</code><code>    <span style="color:#ca7d37">case</span> <span style="color:#dd1144">'Average filter'</span></code><code>        watermarked_image = averageFilter(watermarked_image);</code><code>    <span style="color:#ca7d37">case</span> <span style="color:#dd1144">'JPEG2000 compression'</span></code><code>        watermarked_image = jp2Attack(watermarked_image,param);</code><code>    <span style="color:#ca7d37">case</span> <span style="color:#dd1144">'JPEG compression'</span></code><code>        watermarked_image = jpegAttack(watermarked_image,param);</code><code>    <span style="color:#ca7d37">case</span> <span style="color:#dd1144">'Gaussian low-pass filter'</span></code><code>        watermarked_image = GlowpassFilter(watermarked_image,param);</code><code>    <span style="color:#ca7d37">case</span> <span style="color:#dd1144">'Histogram equalization'</span></code><code>        watermarked_image = histAttack(watermarked_image);</code><code>    <span style="color:#ca7d37">case</span> <span style="color:#dd1144">'Rescaling (0.25)'</span></code><code>        watermarked_image = rescaleAttack(watermarked_image);</code><code>        % please write <span style="color:#ca7d37">this</span> attack code</code><code>    <span style="color:#ca7d37">case</span> <span style="color:#dd1144">'Rescaling (4)'</span></code><code>        watermarked_image = rescale4Attack(watermarked_image);</code><code>        % please write <span style="color:#ca7d37">this</span> attack code</code><code>    <span style="color:#ca7d37">case</span> <span style="color:#dd1144">'Crop attack'</span></code><code>        % please write <span style="color:#ca7d37">this</span> attack code</code><code>        watermarked_image = cropAttack(watermarked_image);</code><code>    <span style="color:#ca7d37">case</span> <span style="color:#dd1144">'Winner filter'</span></code><code>        % please write <span style="color:#ca7d37">this</span> attack code</code><code>    otherwise</code><code>        <span style="color:#dd1144">errordlg</span>(<span style="color:#dd1144">'Please specify attack!'</span>);</code><code>​</code><code>​</code><code>​</code><code>​</code><code>​</code><code>end</code><code>end</code><code>​</code><code>​</code></span></span>

3 Running results

4 reference

[1] Gan Zhichao , Liu Dan . be based on HD and SVD Of DWT Transformed digital image watermarking [J]. Modern information technology , 2022, 6(1):4.

1. Zhang Baosheng . The nature and function of academic evaluation  [J]. Academic research ,2006（2）： 10-15+147.
2. [2] sharpstone .CSSCI And academic evaluation  [J]. Journal of South China Normal University （ Social Sciences Academic Edition ）,2006（4）：135-137.
3.  [3] Zhang Lingling , Lin Meifang . Time —— Scholars under the distribution characteristics of citations P Index research investigate  [J]. Modern intelligence ,2019,39（1）：169-177.
4. [4] Jinbihui ,RONALD R. R Index 、AR Index ：h Index function expansion Supplementary indicators of the exhibition  [J]. Scientific observation ,2007（3）：1-8．
5. [5] Zhuyunxia , Zhu Hui . Research and demonstration of a new academic evaluation index  [J]. JOURNAL OF SOUTHWEST UNIVERSITY FOR NATIONALITIES （ Humanities and Social Sciences ）2012,33（7）：230-233.
6.  [6] Zhu Yihua , Shi Xiaodi .Q The improvement of index and Its Empirical Research  [J]. Intelligence detective Cable ,2020（10）：5-7.

About bloggers ： Good at intelligent optimization algorithms 、 Neural networks predict 、 signal processing 、 Cellular automata 、 The image processing 、 Path planning 、 UAV and other fields Matlab Simulation , relevant matlab Code problems can be exchanged by private letter .

Some theories cite network literature , If there is infringement, contact the blogger to delete .