CN104143172B - A kind of coloured image holographic watermark method based on wavelet de-noising - Google Patents
A kind of coloured image holographic watermark method based on wavelet de-noising Download PDFInfo
- Publication number
- CN104143172B CN104143172B CN201410331437.3A CN201410331437A CN104143172B CN 104143172 B CN104143172 B CN 104143172B CN 201410331437 A CN201410331437 A CN 201410331437A CN 104143172 B CN104143172 B CN 104143172B
- Authority
- CN
- China
- Prior art keywords
- watermark
- obtains
- done
- wavelet
- holographic
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired - Fee Related
Links
Landscapes
- Editing Of Facsimile Originals (AREA)
- Image Processing (AREA)
Abstract
The coloured image holographic watermark method based on wavelet de-noising that the present invention relates to a kind of accurately objectively differentiating watermark information by carrying out wavelet de-noising binary conversion treatment to extraction watermark figure to reach computer.First, forward wavelet transform is carried out to original color image and optical computing holography double random-phase encoding watermark is embedded in, obtain the coloured image containing watermark information；Secondly, to obtained, the coloured image containing watermark information carries out reverse wavelet transform transformation and optical computing holography double random phase decrypts watermark extracting, obtains holographic watermark image；Then, 2-d wavelet noise reduction and binary conversion treatment are done to holographic watermark image, obtains noise reduction watermarking images.Compared with existing digital watermarking algorithm, the present invention is conducive to the analysis and processing of watermarking images, and it is convenient accurately objectively to differentiate that watermark information provides for computer, keeps final watermark extracting figure more acurrate.The present invention can play application value in batch identifies watermark.
Description
Technical field
The present invention relates to a kind of method of digital printing image copyright protection, especially a kind of colours based on wavelet de-noising
Image holographic watermark method, belongs to digital watermark technology field.
Background technology
Digital watermark technology as one of Information Hiding Techniques, save from damage by the digital asset why can become current mainstream
Shield measure is its safety, concealment and robustness.It can not influence former digital content, such as digital printing figure
Under conditions of effective use value of multimedia content such as picture, some identification informations are embedded in digital content, to play protection
The effect of its copyright.
Wavelet technique is one of common Typical Digital watermarking algorithm in recent years, belongs to transform-domain algorithm.Use small echo
The concealment that technology is embedded in watermark is higher, can not visually differentiate, robustness is also high, can effectively improve the safety of watermark.Small echo
Technology is applicable not only to digital watermarking field, also dabbles in many practical applications, for example measurement, estimation, denoising etc., be one very
Effective tool of mathematical analysis.
Calculate holography CGH (Computer-Generated Hologram) technology be it is a kind of establish numerical calculation with show
The encryption emulation mode of generation optically.The hologram made of CGH methods can recording light wave-amplitude and phase comprehensively, have low
The characteristics of noise and high duplication.CGH encrypts emulation technology and simulates entire holographic process using computer technology so that it adds
Real time implementation may be implemented in decryption.
Double random phase encryption and decryption technology is a kind of respectively in spatial domain and frequency domain to hidden image progress random phase volume
The method of code, obtained coded image is the white noise of amplitude and phase random distribution.The process eliminate do not knowing key
In the case of possibility that image is decrypted, confidentiality is higher, and application is wide.But due to the use of Hologram skill
Art, it is possible to find the watermark figure of extraction is unfavorable for the analysis of image and later stage computer is objective sentences there are certain white noise voice phenomenon
Other watermark information, therefore in digital watermark technology field, in order to completely clearly extract watermark figure, use is more comprehensively efficient
Digital watermark method also need further to study.
Invention content
The coloured image holographic watermark method based on wavelet de-noising that the purpose of the present invention is to propose to a kind of is conducive to calculate passenger
It sees and distinguishes watermark information, robustness and integrality for improving watermark can play application value in batch identifies watermark.
To achieve the goals above, present invention employs following scheme：
A kind of coloured image holographic watermark method based on wavelet de-noising, has steps of：(1) to original color image
Forward wavelet transform and optical computing holography double random-phase encoding watermark insertion are carried out, the cromogram containing watermark information is obtained
Picture；(2) to obtained, the coloured image containing watermark information carries out reverse wavelet transform transformation and the holographic double random phases of optical computing
Position decryption watermark extracting, obtains holographic watermark image；(3) 2-d wavelet noise reduction and binary conversion treatment are done to holographic watermark image,
Obtain noise reduction watermarking images.
The present invention has the advantage that compared with existing digital watermark：This method Blind extracting, using wavelet transformation with it is double
Under the premise of random phase optical computing encryption and transform domain insertion watermark, wavelet de-noising two-value is carried out when extracting watermarking images
The processing of change.Safe, good concealment is conducive to computer accurately objective differentiation watermark information.Compared with existing digital watermark,
The technology not only significantly improves the robustness of watermark, but also can allow the better automatic identification watermark information of computer, batch
Application value can be played in amount identification watermark.
Description of the drawings
Fig. 1 is the insertion flow chart of the coloured image holographic watermark based on wavelet de-noising of the application.
Fig. 2 is the substep design sketch of the insertion watermark information of flow shown in FIG. 1.
Fig. 3 is the extraction flow chart of the coloured image holographic watermark based on wavelet de-noising of the application.
Fig. 4 is the design sketch of the wavelet de-noising binaryzation in flow shown in Fig. 3.
Fig. 5 is Attack Digital Watermarking part of detecting simulated effect figure.
Specific implementation mode
Below in conjunction with attached drawing to a kind of coloured image holographic watermark method based on wavelet de-noising according to the present invention
Preferred embodiment is made to illustrate in detail, but the present invention is not limited only to the embodiment.In order to make the public have thoroughly the present invention
Solution, has carried out concrete details explanation in preferred embodiment below.
The preferred embodiment of the present invention under 64 Window7 operating systems, selection standard test chart Lena figure (as
Element is 512 × 512) and watermark figure (pixel is 128 × 128), select the channels independence based on human eye visual perception preferable
Lab color spaces make carrier, are simulated and are obtained using MATLAB2011b softwares.
A kind of coloured image holographic watermark method based on wavelet de-noising of the present invention includes watermark insertion and watermark extracting
Two parts carry out forward wavelet transform to original color image and optical computing holography double random-phase encoding watermark are embedding first
Enter, obtains the coloured image containing watermark information；Then, to containing watermark information coloured image carry out reverse wavelet transform transformation and
Optical computing holography double random phase decrypts watermark extracting, obtains holographic watermark image；Finally, two dimension is done to holographic watermark image
Wavelet de-noising and binary conversion treatment obtain objective calculating watermark figure, i.e. noise reduction watermarking images.
As shown in Figure 1, the coloured image holographic watermark insertion flow based on wavelet de-noising includes having the following steps：
(1) color space conversion is done (shown in such as Fig. 2 (a)) to original color image, obtains a lightness or brightness figure layer
L1 (shown in such as Fig. 2 (b)) and two color information figure layer A and B (such as Fig. 2 (c) and Fig. 2 (d) are shown), retain two color figure layers
Information A and B；(2) two level wavelet transformation is done to lightness or luminance information figure layer L1, obtains high-frequency sub-band HH2；(3) to initial condition
Watermark image does double random-phase encoding (shown in such as Fig. 2 (g)), that is, calculates holographic transformation, obtain encrypted encrypted hologram,
Referred to as holographic watermark figure (such as Fig. 2 (h))；(4) obtained holographic watermark figure is embedded in two level high frequency obtained in (2) step
Band HH2 obtains the high-frequency sub-band HH2 ' of embedded Hologram watermark information, since human eye is not high to high-frequency information susceptibility, entirely
Cease watermark figure has high noise immunity again, so the L1 layers obtained after embedded two level wavelet transformation after watermark information is weighted
Two level high-frequency sub-band HH2 ensures robustness and noise immunity with this, while also increasing the invisibility of vision.(5) to high-frequency sub-band
HH2 ' does reverse two level haar wavelet transform, obtains the figure layer L1 ' containing watermark information (shown in such as Fig. 2 (e))；(6) by upper one
Two color the map data mining platforms A and B for walking obtained figure layer L1 ' and retaining in (1) step merge, and obtain containing holographic watermark information
Coloured image (such as Fig. 2 (f) shown in).
Wherein, the two level wavelet transformation in step (2) specifically comprises the following steps：(21) first to lightness or luminance information figure
Layer L1 does a haar wavelet transform, obtains four L1 layers of primary frequency subband, this four primary frequency subbands include one
Low frequency sub-band LL1 and three high-frequency sub-bands LH1, HL1 and HH1；(22) haar wavelet transform is done to the low frequency sub-band LL1 again,
Obtain four L1 layers of Second Level Frequency subband：LL2, LH1, HL2 and HH2.
It is the substep design sketch that watermark information is embedded according to the flow chart of Fig. 1 referring now to Fig. 2, Fig. 2.
As shown in Fig. 2, Fig. 2 (a) is original color image, Fig. 2 (b) is obtained lightness figure layer after color space conversion
L1, Fig. 2 (c) and Fig. 2 (d) are respectively color information the figure layer A and B obtained after color space conversion.Fig. 2 (g) is original watermark
Image generates the holographic watermark figure of Fig. 2 (h) after doing double random phase modulation calculating Hologram to it.Fig. 2 (e) is embedded water
The lightness information figure layer L1 ' obtained through two level wavelet inverse transformation after print.Finally, by the lightness information figure layer L1 ' of embedded watermark with
Colouring information the figure layer A and B previously retained carries out multi-layer image merging treatment, obtains the colored number containing holographic watermark of Fig. 2 (f)
Word image.
It is the extraction flow chart of the coloured image holographic watermark based on wavelet de-noising of the application referring now to Fig. 3, Fig. 3.
As shown in figure 3, the coloured image holographic watermark extraction flow based on wavelet de-noising includes having the following steps：
(1) color is done to targeted color image (the aqueous printed color picture that method flow diagram i.e. according to figure 1 generates)
Space is converted, and a lightness or brightness figure layer L2 and two color information figure layers A ' and B ' are obtained；(2) to lightness or luminance information
Figure layer L2 is two level wavelet transformation, high-frequency sub-band HH4；(3) double random phase modulation decryption is done to the HH4 obtained in previous step,
It calculates holography, obtains holographic watermark image (shown in such as Fig. 4 (a))；(4) 2-d wavelet noise reduction and two are done to holographic watermark image
Value, obtain it is objective calculate watermark figure, i.e. noise reduction watermarking images (shown in such as Fig. 4 (c)).
Wherein step (2) is done two level wavelet transformation to lightness or luminance information figure layer L2 and is specifically comprised the following steps：(21)
Haar wavelet transform is done to lightness or luminance information figure layer L2, obtains four L2 layers of primary frequency subband：One low frequency sub-band
LL3 and three high-frequency sub-band LH3, HL3 and HH3；(22) it tries again haar wavelet transform, obtains to the low frequency sub-band LL3
Four L2 layers of Second Level Frequency subband：LL4, LH4, HL4 and HH4.
Wherein step (4) does 2-d wavelet noise reduction to holographic watermark image and binaryzation is as follows：
(41) the Birge-Massart algorithms in wavelet technique are used, threshold deniosing is done to the watermark hologram extracted
Processing.It is calculated using wavelet coefficient selection small echo Birge-Massart algorithms, after the threshold value for obtaining one-dimensional wavelet transform
Carry out noise reduction.Formula is as follows：
Crit (t)=| sum (c (k)2, k | t)+2 × σ2×t×(α+log(n/t)) (1)
In formula (1), the threshold value that cirt (t) is the t for obtaining minimum value judges formula.Wherein, σ is the Gauss white noise of zero-mean
The standard deviation of sound；α is adjustable parameter, is a real number more than 1, generally takes 2；C (k) is arranged from big to small by absolute value
The big wavelet packet coefficient of kth；N is the number of coefficient.In MATLAB emulation experiments, the wbmpen of wavelet toolbox can be called
Function and wdcbm functions obtain threshold value and noise reduction.
(42) binary conversion treatment is done to the holographic watermark image after noise reduction process.
It is wavelet de-noising and the design sketch of binaryzation referring now to Fig. 4, Fig. 4.
As shown in figure 4, Fig. 4 (a) is extraction watermark artwork, i.e., the holographic watermark image in method flow diagram shown in Fig. 3.
When Fig. 4 (b) is that the parameter alpha in formula (1) takes 1.2, the design sketch after two level wavelet de-noising.It, can be with due to the use of Hologram
Find out that white noise is very big.Fig. 4 (c) is the design sketch after noise reduction binaryzation, and watermarking images are high-visible.
It is Attack Digital Watermarking part of detecting simulated effect figure referring now to Fig. 5, Fig. 5.
As shown in figure 5, Fig. 5 (a) be using quality factor parameter Quality (0.2) jpeg formats to having been inserted into watermark figure
After being preserved as compression, then read-in programme detects the test result of watermark.Fig. 5 (b) is using incremental compression ratio parameter
Test result when Compression Ratio (80%).Fspecial functions come simulation filter attack, institute in Calling MATLAB
There is filtering to be all made of the Filtering Template of 3 × 3 sizes, obtains the mean filter test result such as Fig. 5 (c).Fig. 5 (f) is that Gauss is low
The design sketch of pass filter, template standard difference Sigma take its default value 0.5.Fig. 5 (d) is the design sketch of sharp filtering, Laplce
The edge factor is 0.2.Fig. 5 (e) is the design sketch of motion blur, and θ takes 0, LEN to take 9, indicates that image is moved with 0 ° in the counterclockwise direction
9 pixels are moved；It is tested respectively with four kinds of Gaussian noise, salt-pepper noise, speckle noise and poisson noise attacked by noise.Its
The test result that middle Gaussian noise intensity is 0.001~0.01, Fig. 5 (g) when being high intensity 0.01.Salt-pepper noise and speckle noise
Strength range be 0.01~0.1, Fig. 5 (h) and Fig. 5 (i) be respectively the survey of salt-pepper noise and speckle noise in high intensity 0.1
Test result.Fig. 5 (j) is the test result of poisson noise.Shearing attack is carried out to the original digital image containing watermark, cuts range
It is test result when cutting 50% from 10%~50%, Fig. 5 (k).
Referring now to table 1, table 1 is Attack Digital Watermarking test data table.
Table 1
As shown in table 1, PSNR is the Y-PSNR of Attack Digital Watermarking figure and artwork, and PSNR is smaller, illustrates that distortion is bigger.NC
Value is watermarking images and raw water watermark image, and closer to 1, surface extraction effect is better.The algorithm is in Lab it can be seen from table
In color space, for compression, filtering, fuzzy and each noise like, NC values all 0.9 or more, have preferable robustness.And it is right
In shearing, when PSNR only has 1.252, that is, when cutting 50%, NC values still keep 0.8 or more, and watermark extracting figure is also clear
It can be seen that.
Feature, entirety, feature or grouping with the description of certain aspects of the present disclosure, embodiment or example collaboration should be by
It is interpreted as can be applied in other aspect, embodiment or examples described herein, unless not compatible therewith.This specification
All features and/or disclosed any method or stream disclosed in (including any the appended claims, abstract and attached drawing)
Overall Steps in journey can be merged with any combinations, except in non-combined at least partly feature and/or step be mutual exclusion
's.The present invention is not limited to any details of any aforementioned embodiments.It is (including any additional that the present invention expands to this specification
Claims, abstract and attached drawing) disclosed in the either combination of novel feature or the extension of any one novel feature
To the step combination of any one novel step or novelty of disclosed any method.
Claims (1)
1. a kind of coloured image holographic watermark method based on wavelet de-noising, which is characterized in that this approach includes the following steps：
(1) forward wavelet transform is carried out to original color image and optical computing holography double random-phase encoding watermark is embedded in, obtained
To the coloured image containing watermark information；
(2) to obtained, the coloured image containing watermark information carries out reverse wavelet transform transformation and the holographic double random phases of optical computing
Position decryption watermark extracting, obtains holographic watermark image；
(3) 2-d wavelet noise reduction and binary conversion treatment are done to holographic watermark image, obtains noise reduction watermarking images；
The step (1) further comprises following steps：
(11) color space conversion is done to original color image, obtains a lightness or brightness figure layer L1 and two color information figures
Layer A and B；
(12) two level wavelet transformation is done to lightness or luminance information figure layer L1, obtains high-frequency sub-band HH2；
(13) double random-phase encoding is done to original watermark image, that is, calculates holographic transformation, obtains encrypted holographic watermark figure；
(14) obtained holographic watermark figure is embedded in the two level high-frequency sub-band HH2 obtained in (12) step, obtains embedded holography
The high-frequency sub-band HH2 ' of encrypted watermark information；
(15) reverse two level haar wavelet transform is done to high-frequency sub-band HH2 ', obtains the figure layer L1 ' containing watermark information, and incite somebody to action
To figure layer L1 ' and (11) step in obtained two color information figure layer A and B merge, obtain the colour containing watermark information
Image；
(4) various attack tests are carried out to the noise reduction watermarking images, the attack test includes：
Wherein step (12) further comprises following steps：
Haar wavelet transform is done to lightness or luminance information figure layer L1, obtains four L1 layers of primary frequency subband, this four
Grade frequency subband includes a low frequency sub-band LL1 and three high-frequency sub-bands LH1, HL1 and HH1；
Haar wavelet transform is done to the low frequency sub-band LL1, obtains four L1 layers of Second Level Frequency subband：LL2, LH2, HL2 and
HH2；
The step (2) further comprises following steps：
(21) color space transformation is done to the coloured image containing watermark information, obtains a lightness or brightness figure layer L2 and two
A color information figure layer A ' and B '；
(22) two level wavelet transformation is done to lightness or luminance information figure layer L2, obtains high-frequency sub-band HH4；
(23) double random phase modulation decryption is done to the HH4 obtained in previous step, that is, calculates holography, obtains holographic watermark image；
(24) 2-d wavelet noise reduction and binaryzation are done to holographic watermark image, obtains noise reduction watermarking images；
The step (22) further comprises following steps：
Haar wavelet transform is done to lightness or luminance information figure layer L2, obtains four L2 layers of primary frequency subband：One low frequency
Subband LL3 and three high-frequency sub-bands LH3, HL3 and HH3；
Haar wavelet transform is done to the low frequency sub-band LL3, obtains four L2 layers of Second Level Frequency subband：LL4, LH4, HL4 and
HH4；
The step (3) further comprises following steps：
(31) the Birge-Massart algorithms in wavelet technique are used, threshold deniosing processing is done to the holographic watermark figure extracted；
(32) binary conversion treatment is done to the holographic watermark image after noise reduction process；
Birge-Massart algorithmic formulas employed in the step (31) are as follows：
Crit (t)=Isum (c (k)2, k | t)+2 × σ2×t×(α+log(n/t))
Wherein, the threshold value that cirt (t) is the t for obtaining minimum value judges formula；σ is the standard deviation of the white Gaussian noise of zero-mean
Difference；α is adjustable parameter, is a real number more than 1；C (k) is to arrange the big wavelet packet coefficient of kth from big to small by absolute value；
N is the number of coefficient.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410331437.3A CN104143172B (en) | 2014-07-11 | 2014-07-11 | A kind of coloured image holographic watermark method based on wavelet de-noising |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410331437.3A CN104143172B (en) | 2014-07-11 | 2014-07-11 | A kind of coloured image holographic watermark method based on wavelet de-noising |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104143172A CN104143172A (en) | 2014-11-12 |
| CN104143172B true CN104143172B (en) | 2018-09-18 |
Family
ID=51852340
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410331437.3A Expired - Fee Related CN104143172B (en) | 2014-07-11 | 2014-07-11 | A kind of coloured image holographic watermark method based on wavelet de-noising |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104143172B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105023237A (en) * | 2015-07-17 | 2015-11-04 | 上海理工大学 | Method for improving concealment performance of image digital watermarks |
| CN105931174A (en) * | 2016-04-12 | 2016-09-07 | 上海理工大学 | High-robustness digital watermarking method making color QR code embedded in color image |
| CN110400247B (en) * | 2019-07-16 | 2023-06-02 | 上海理工大学 | Printing, printing and shooting resistant semi-blind watermarking method based on color image |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102592257B (en) * | 2012-01-04 | 2013-07-24 | 上海理工大学 | Cyan, magenta, yellow and black (CMYK) color space-based optical holographic watermark algorithm |
| CN103077495A (en) * | 2013-02-01 | 2013-05-01 | 上海理工大学 | Digital holography watermarking algorithm based on QR (quick response) code |
| CN103366333A (en) * | 2013-07-09 | 2013-10-23 | 上海理工大学 | Standard color space-based color digital holographic watermarking algorithm |
-
2014
- 2014-07-11 CN CN201410331437.3A patent/CN104143172B/en not_active Expired - Fee Related
Also Published As
| Publication number | Publication date |
|---|---|
| CN104143172A (en) | 2014-11-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Yu | Attention based data hiding with generative adversarial networks | |
| US7720253B2 (en) | Visual perception model for hi-fidelity image watermarking | |
| CN108564519B (en) | Color image digital watermarking method based on multiple transform domains | |
| Malonia et al. | Digital image watermarking using discrete wavelet transform and arithmetic progression technique | |
| CN106570813A (en) | Holographic digital watermark embedding method, extraction method and device | |
| CN104143172B (en) | A kind of coloured image holographic watermark method based on wavelet de-noising | |
| CN106131711A (en) | Robust hidden watermark embedding towards 3D high-definition digital video and extracting method | |
| Pithiya et al. | DWT based digital image watermarking, de-watermarking & authentication | |
| Dhaou et al. | An Efficient Group of Pictures Decomposition based Watermarking for Anaglyph 3D Video. | |
| CN105023237A (en) | Method for improving concealment performance of image digital watermarks | |
| Hoshi et al. | A robust watermark algorithm for copyright protection by using 5-level DWT and two logos | |
| Khare et al. | Secure and robust image watermarking scheme using homomorphic transform, SVD and Arnold transform in RDWT domain | |
| Srinivas et al. | A new video watermarking using redundant discrete wavelet in singular value decomposition domain | |
| Singha et al. | Transform domain digital watermarking with multiple images as watermarks | |
| Kude et al. | Skintone detection based steganography using wavelet transform | |
| Mohammed et al. | Digital image watermarking, analysis of current methods | |
| Jabbar et al. | Compare between DCT and DWT for digital watermarking in color image | |
| Liang et al. | Ridgelet-based robust and perceptual watermarking for images | |
| Wang | Robust watermarking in wavelet domain based on chaotic scrambling | |
| Barkat et al. | Time-frequency and time-scale-based fragile watermarking methods for image authentication | |
| Ramaiya et al. | Digital security using watermarking techniques via discrete wavelet transform | |
| Onoja et al. | Robust Watermarking Techniques for the Authentication and Copyright Protection of Digital Images: A Survey | |
| Liu et al. | Robustness comparison of color image watermarking schemes in uniform and non-uniform color spaces | |
| Kumari et al. | Robust image watermarking algorithm DWT-SVD-AES for telemedicine application | |
| CN117115110A (en) | Morphology operation-based accurate image tampering evidence obtaining method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20180918 Termination date: 20210711 |