CN103559676A - Method for preventing color image blind watermarking from being printed and scanned based on DCT coefficient statistical property - Google Patents

Method for preventing color image blind watermarking from being printed and scanned based on DCT coefficient statistical property Download PDF

Info

Publication number
CN103559676A
CN103559676A CN201310473762.9A CN201310473762A CN103559676A CN 103559676 A CN103559676 A CN 103559676A CN 201310473762 A CN201310473762 A CN 201310473762A CN 103559676 A CN103559676 A CN 103559676A
Authority
CN
China
Prior art keywords
coefficient
watermark
watermarking
group
image
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.)
Pending
Application number
CN201310473762.9A
Other languages
Chinese (zh)
Inventor
王泉
罗楠
杨鹏飞
潘蓉
郑海红
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Xidian University
Original Assignee
Xidian University
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Xidian University filed Critical Xidian University
Priority to CN201310473762.9A priority Critical patent/CN103559676A/en
Publication of CN103559676A publication Critical patent/CN103559676A/en
Pending legal-status Critical Current

Links

Images

Abstract

The invention discloses a method for preventing a color image blind watermarking from being printed and scanned based on a DCT coefficient statistical property. The method mainly solves the problems that in the traditional algorithm, the watermarking embedding capacity is low, robustness is poor, non-blind extraction mode is adopted, and the like. According to the method, the character that the mean value of intermediate frequency coefficients of image blocks in DCT transformation in statistics is approximately equal to zero, positive and negative offset is carried out on the coefficients according to embedded watermarking positions, self-adaptation adjustment is carried out on watermarking embedding strength according to image textural features, and therefore watermarking information can be embedded in a blue channel of a color image. In the process of watermarking extraction, only block DCT transformation on blue elements of the image is needed, the mean value of the corresponding intermediate frequency coefficients is counted and compared with zero, and therefore information of the embedded watermarking can be obtained. Compared with the traditional algorithm for preventing the image watermarking from being printed and scanned, the method has the advantages that under the precondition that high invisibility and high robustness of the watermarking are guaranteed, the watermarking embedding capacity is improved, blind extraction is achieved, and the embedded watermarking information is strong for printing, copying and scanning channels.

Description

Anti printing and scanning coloured image blind watermark method based on DCT coefficients statistics characteristic
Technical field
The invention belongs to digital watermark technology field, relate to embedding and the extraction of Color digital watermarking information.The blind watermark method that specifically a kind of coloured image anti printing and scanning based on DCT coefficients statistics characteristic is attacked, the method is guaranteeing under the prerequisite of strong robustness and large embedding capacity, obtained good visual effect, and the watermark information embedding is healthy and strong to printing scan channel.
Background technology
In the last few years, progress along with quality digital input and output technology, and being widely used of the equipment such as digital camera, video recorder, scanner and printer, particularly high precision color laser printer, scanner is universal, for people's live and work has brought many facilities, but also make distorting or forge and becoming very easy of various certificates, the packing of product, security, check and proof document simultaneously.Therefore, these images of producing with printing process, the copyright protection of document have just become problem anxious to be resolved, make antiforge technology seem extremely important.
Digital image watermarking technology can address this problem well, can reach good effect, can effectively reduce costs again.Digital image watermarking technology will have image or the text mark (being digital watermarking) of certain sense; utilize the method for Information hiding to be embedded in digital picture product; in order to prove the entitlement of creator to its works; and can be used as the foundation of evaluation, the illegal infringement of prosecution; simultaneously by the determination and analysis of watermark being carried out to the complete reliable of guarantee information, thereby become intellectual property protection and the false proof effective means of digital multimedia.Compare with traditional anti-counterfeiting technology, digital figure watermark is false proof has the features such as safe, detection speed is fast, cost is low, this technology is applied to antiforge and will has very large prospect.
At present, according to watermark implementation procedure, watermarking algorithm is broadly divided into two classes: spatial domain algorithm and transform-domain algorithm.
1, space watermark algorithm
Space watermark algorithm carrys out embed watermark by the pixel value of direct modification original image, and wherein most typical is least significant bit (LSB) method fLSB, least significant bit), the Patchwork algorithm etc. based on statistics; Yet, because space watermark embeds algorithm, be directly to revise pixel value to carry out embed watermark, its hiding watermark information is difficult to ensure and deposits under stronger signal disposition, especially the far-reaching hard copy to picture quality as print-scanning is attacked, and above-mentioned spatial domain algorithm is difficult to successful Application.
2. transform domain watermarking algorithm
This class algorithm is to realize in transform domain, by revising some conversion coefficient, carrys out embed watermark.This class technology-as be part based on conventional or global image conversion, compare with spatial domain watermark algorithm, transform domain watermarking algorithm has following advantage:
(1) watermark signal that transform domain embeds can be distributed in all pixels in spatial domain, is conducive to guarantee the invisibility of watermark, increases the anti-attack ability of watermarking algorithm simultaneously;
(2) in transform domain, can be better in conjunction with HVSf human visual system) masking characteristics, make embedment strength, embedded location, the self-adaptation adjustment along with the difference of image-carrier data each several part characteristic of embedding quantity of information of watermark, owing to considering the local feature of initial carrier, finally realize the best compromise of invisibility and robustness.
(3) transform domain method can be compatible with existing method for compressing image, can in compression, complete watermark encoder, strengthens the incompressible attacking ability of watermaking system, and simplify watermark embed process.
Just because of these several large advantages, numerous scholars are chosen in transform domain embed watermark, and the image conversion of main flow has discrete cosine transform (DCT), wavelet transform (DWT) and discrete Fourier transform (DFT) (DFT) etc.
Since anti printing and scanning digital figure watermark algorithm proposes, numerous scholars have carried out large quantity research, but because print scanned process is an actual applied environment, the suffered attack of watermark signal embedding is not one or more the different simple combination of attacking in simple emulation experiment, but the result of Various Complex signal synthesis effect.Most algorithm is just studied for a certain attack in print scanned process, and the watermark effect of extraction is not good; And embedded watermark signal mostly is a small amount of bit sequence, or need original image just can propose watermark information, from actual application level, also differ greatly.
Summary of the invention
The object of the invention is to the deficiencies such as, poor robustness low for the embedding capacity existing in existing anti printing and scanning watermarking algorithm, non-blind watermark extracting pattern, propose a kind of coloured image blind watermark method of the anti printing and scanning based on DCT coefficients statistics characteristic.The watermark information that has guaranteed on the one hand embedded images can also correctly be extracted after printing, copying and scan under blind watermark extracting pattern; Realized on the other hand the simple and safety of the high embedding capacity of watermark, high invisibility and watermark embedding and extraction system.
Realizing technical thought of the present invention is: the characteristic of utilizing mean approximation that the intermediate frequency coefficient of image block dct transform presents in statistics to equal 0, according to the watermark bit embedding, coefficient is applied to positive and negative skew, and regulate watermark embed strength according to image texture characteristic self-adaptation, realize embed watermark information in coloured image blue channel; When watermark extracting, only need carry out block DCT transform to image blue component, add up the average of corresponding intermediate frequency coefficient, and with 0 relatively, just can obtain the watermark information of embedding.Specifically comprise the following steps:
(1) watermark of coloured image embeds
A. watermark information pre-service, carries out scramble by binary bitmap, by row major, is arranged as one-dimensional sequence, obtains 01 watermark sequence to be embedded;
B. extract the blue component of coloured image, blue component is carried out to block DCT transform operation, and calculate the texture mask value of each pixel of blue component;
C. adding up the intermediate frequency coefficient of block DCT transform, is one group by the coefficients statistics of every middle same position, can obtain a plurality of coefficient sets altogether, and group number scale is N; Again coefficient sets is divided, each original coefficient sets is split into 4 subgroups, can obtain 4N group altogether;
D. to every group of coefficient, according to the watermark bit embedding, be 1 or 0, apply the skew of forward or negative sense, make its average be greater than 0 or be less than 0, the size of skew is determined by the texture mask value of the pixel corresponding with coefficient positions, embeds a watermark in every group of coefficient;
E. each DCT coefficient block is implemented to inverse transformation, the coloured image after being embedded, completes watermark embed process.
(2) watermark extracting of coloured image
A. extract the blue component of coloured image to be detected, it is carried out to block DCT transform operation;
B. adding up the intermediate frequency coefficient of block DCT transform, is one group by the coefficients statistics of every middle same position, and every group of coefficient is split into 4 subgroups;
C. calculate the average of every group of coefficient, and 0 compare, 01 watermark sequence that obtains extracting, then sequence is rearranged to two dimensional image, and carry out disorderly conversion of inverted, obtain the binary bitmap of extracting.
(3) the present invention compares and has the following advantages with traditional anti printing and scanning Watermarking of Color:
A. the present invention carrys out embed watermark according to the statistical property of the intermediate frequency coefficient of image block dct transform, this specific character at image after attacks of print_scan, still keep good, at this print scanned " invariant " upper embed watermark, can obtain good anti printing and scanning robustness.
B. the present invention carries out block DCT transform with suitable size to image, statistics intermediate frequency coefficient, and every group of coefficient further divided, to obtain more coefficient sets, finally at every group, embed a watermark, improved watermark embedding capacity; And every watermark is all distributed in entire image, thereby improved the anti-attack ability of algorithm.
C. in the present invention, adopt texture mask model as the measurement factor of watermark embed strength, texture mask has embodied the masking characteristics of image local texture, texture is abundanter, and its acceptable embedment strength is discovered by human eye more greatly and not, namely can guarantee good watermark invisibility; Vice versa.
D. the present invention is when extracting watermark information, only need to treat detected image carries out block DCT transform, then calculate intermediate frequency coefficient group average and with 0 relatively, can obtain watermark information, and do not need to provide the participation of original image, therefore realize the blind extraction pattern of watermark completely, there is the practical value of range of application and Geng Gao widely.
Accompanying drawing explanation
By describing in more detail exemplary embodiment of the present invention with reference to accompanying drawing, above and other aspect of the present invention and advantage will become and more be readily clear of, in the accompanying drawings:
Fig. 1 is the schematic flow sheet that the present invention adopts DCT coefficients statistics characteristic embed watermark in coloured image;
Fig. 2 is the schematic diagram of DCT intermediate frequency coefficient;
Fig. 3 is that the present invention adopts DCT coefficients statistics characteristic in coloured image, to extract the schematic flow sheet of watermark.
Embodiment:
Below in conjunction with the accompanying drawing in the embodiment of the present invention, the technical scheme in the embodiment of the present invention is carried out to clear, complete description, obviously, described embodiment is only a part of embodiment of the present invention, rather than whole embodiment.Embodiment based in the present invention, those of ordinary skills, not making the every other embodiment obtaining under creative work prerequisite, belong to the scope of protection of the invention.
Embodiment 1
The watermark of coloured image embeds concrete implementation step contrast Fig. 1, is described below:
Step 1, watermark information pre-service.
The binary bitmap that will embed is carried out Arnold scramble, and the image after scramble is arranged as to one-dimensional sequence by row major, obtains 01 watermark sequence to be embedded.
Step 2, texture mask calculates.
Obtain the blue component figure of coloured image, calculate the texture mask value of blue component each point, be designated as m (i, j).
m ( i , j ) = | x ( i , j ) - x _ ( i , j ) | ,
x _ ( i , j ) = 1 ( 2 L + 1 ) 2 Σ k = - L L Σ l = - L L x ( i + k , j + l )
Wherein, x (i, j) is the pixel value that position (i, j) is located,
Figure BDA0000393944130000043
be the mean value of pixel in the sliding window centered by (i, j), m (i, j) is the texture mask value that (i, j) locates pixel, (2L+1) 2number of pixels, here L=1 in the sliding window of representative.
Step 3, minute block operations and coefficients statistics.
Blue component figure to obtaining, carries out block DCT transform with 16 * 16 sizes, collects intermediate frequency coefficient, obtains intermediate frequency coefficient set, is designated as C.
C={(u，v,k)：l≤u+v≤r，0≤u，v≤15，1≤k≤k}
Wherein, u, v is DCT coefficient subscript index, and k is piece index, and K is piecemeal sum.
Get l=6 and r=26 defines intermediate frequency coefficient, every can comprise 225 DCT intermediate frequency coefficients (as shown in Figure 2) like this; The intermediate frequency coefficient of every interior same position is added up, can be obtained 225 groups of coefficients, every group of C icomprise K coefficient; The coefficient of K in every group be take and 4 as interval, extracted, thereby be split into 4 subgroup C i1, C i2, C i3, C i4, finally obtain 900 coefficient sets, every group comprises K/4 coefficient.
Step 4, deviation ratio average.
To every group of coefficient, according to the watermark bit embedding, be 1 or 0, apply the skew of forward or negative sense, make its average be greater than 0 or be less than 0, the size of skew is determined by the texture mask value of the pixel corresponding with coefficient positions, embeds a watermark in every group of coefficient.
y u , v k = x u , v k + m u , v k * w i , ( u , v , k ) ∈ C i x u , v k , ( u , v , k ) ∈ C i , ∀ i
Wherein,
Figure BDA0000393944130000052
the DCT coefficient that in piece k, (u, v) locates,
Figure BDA0000393944130000053
the texture mask value of image k piece (u, v) coordinate place pixel,
Figure BDA0000393944130000054
it is the DCT coefficient of embed watermark.W also will illustrate
Step 5, the image of acquisition embed watermark.
Each piece DCT coefficient of revising is done to DCT inverse transformation, obtain the blue component figure after embed watermark, it and other channel data are merged, obtain the coloured image of embed watermark, the watermark that has completed coloured image embeds.
Embodiment 2
The concrete implementation step contrast of watermark extracting Fig. 3 of coloured image, is described below:
Step 1, obtains coloured image blue component.
Step 2, minute block operations and coefficients statistics.
With Color digital watermarking, embed step 3.
Step 3, watermark sequence extracts.
Calculate the average of every group of coefficient, itself and 0 value are compared, if be greater than 0, watermark bit is 1; If be less than 0, watermark bit is 0, so obtains one dimension 01 watermark sequence extracting.
Step 4, binary bitmap generates.
The watermark sequence that step 3 is obtained is rearranged to two dimensional image, and carries out disorderly conversion of inverted, obtains the binary bitmap of extracting.
The foregoing is only embodiments of the invention, be not limited to the present invention.The present invention can have various suitable changes and variation.All any modifications of doing within the spirit and principles in the present invention, be equal to replacement, improvement etc., within protection scope of the present invention all should be included in.

Claims (3)

1. the anti printing and scanning coloured image blind watermark method based on DCT coefficients statistics characteristic, is characterized in that, comprises the steps:
A: the watermark of coloured image embeds;
A1: watermarking images pre-service, binary bitmap is carried out to scramble, by row major, be arranged as one-dimensional sequence, obtain 01 watermark sequence to be embedded;
A2: extract the blue component of coloured image, blue component is carried out to block DCT transform operation, and calculate the texture mask value of each pixel of blue component;
A3: the intermediate frequency coefficient of statistics block DCT transform is one group by the coefficients statistics of every middle same position, can obtain a plurality of coefficient sets altogether, and number is designated as N; Again coefficient sets is divided, each original coefficient sets is split into 4 subgroups, can obtain 4N group altogether;
A4: to every group of coefficient, be 1 or 0 according to the watermark bit embedding, apply the skew of forward or negative sense, make its average be greater than 0 or be less than 0, the size of skew is determined by the texture mask value of the pixel corresponding with coefficient positions, embeds a watermark in every group of coefficient;
A5: each DCT coefficient block is implemented to inverse transformation, obtain the coloured image after embed watermark, complete watermark embed process;
B: the watermark extracting of coloured image;
B1: extract the blue component of coloured image to be detected, it is carried out to block DCT transform operation;
B2: the intermediate frequency coefficient of statistics block DCT transform, is one group by the coefficients statistics of every middle same position, and every group of coefficient is split into 4 subgroups;
B3: calculate the average of every group of coefficient, and with 0 compare, obtain 01 watermark sequence that embeds, then sequence be rearranged to two dimensional image, and carry out disorderly conversion of inverted, obtain the binary bitmap of extracting.
2. the anti printing and scanning coloured image blind watermark method based on DCT coefficients statistics characteristic according to claim 1, it is characterized in that, described block DCT transform operation, that blue component is carried out to block DCT transform with 16 * 16 sizes, 225 coefficients in choosing every are as intermediate frequency coefficient, the intermediate frequency coefficient of every middle same position is added up, can be obtained 225 groups of coefficients; Then the coefficient in every group be take and 4 as interval, extract, thereby each coefficient sets is just split into 4 subgroups, altogether obtains 900 coefficient sets, embeds a watermark in every group of coefficient, has improved watermark embedding capacity.
3. the anti printing and scanning coloured image blind watermark method based on DCT coefficients statistics characteristic according to claim 1, it is characterized in that, the bias size of coefficient is to be determined by the texture mask value of the pixel corresponding with coefficient positions, it is also large that the coefficient of region that texture mask is large is revised amplitude, otherwise also little.
CN201310473762.9A 2013-10-11 2013-10-11 Method for preventing color image blind watermarking from being printed and scanned based on DCT coefficient statistical property Pending CN103559676A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201310473762.9A CN103559676A (en) 2013-10-11 2013-10-11 Method for preventing color image blind watermarking from being printed and scanned based on DCT coefficient statistical property

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201310473762.9A CN103559676A (en) 2013-10-11 2013-10-11 Method for preventing color image blind watermarking from being printed and scanned based on DCT coefficient statistical property

Publications (1)

Publication Number Publication Date
CN103559676A true CN103559676A (en) 2014-02-05

Family

ID=50013916

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201310473762.9A Pending CN103559676A (en) 2013-10-11 2013-10-11 Method for preventing color image blind watermarking from being printed and scanned based on DCT coefficient statistical property

Country Status (1)

Country Link
CN (1) CN103559676A (en)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106373077A (en) * 2016-08-19 2017-02-01 西安电子科技大学 Relative ratio-based digital watermarking method
CN106373079A (en) * 2016-08-31 2017-02-01 南京师范大学 Android photo position and time digital watermark embedding and extraction method
CN108765256A (en) * 2018-07-10 2018-11-06 天津工业大学 Data waterprint embedded method based on a kind of dct transform that human vision is covered
CN112017097A (en) * 2020-08-27 2020-12-01 豪威科技(武汉)有限公司 Watermark adding method, device, terminal and computer readable storage medium

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101661608A (en) * 2009-08-31 2010-03-03 重庆大学 Diginal image blind watermark adding, extracting and removing method suitable for internet
CN101739657A (en) * 2009-11-06 2010-06-16 上海第二工业大学 DCT domain-based color image digital watermark embedding method and DCT domain-based color image digital watermark extracting method
JP2010245876A (en) * 2009-04-07 2010-10-28 Nippon Telegr & Teleph Corp <Ntt> Method and device for embedding electronic watermark for detecting falsification and program, and falsification detection and device using electronic watermark and program

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2010245876A (en) * 2009-04-07 2010-10-28 Nippon Telegr & Teleph Corp <Ntt> Method and device for embedding electronic watermark for detecting falsification and program, and falsification detection and device using electronic watermark and program
CN101661608A (en) * 2009-08-31 2010-03-03 重庆大学 Diginal image blind watermark adding, extracting and removing method suitable for internet
CN101739657A (en) * 2009-11-06 2010-06-16 上海第二工业大学 DCT domain-based color image digital watermark embedding method and DCT domain-based color image digital watermark extracting method

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
罗楠: "抗硬拷贝攻击彩色图像水印算法研究", 《中国优秀硕士学位论文全文数据库》 *

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106373077A (en) * 2016-08-19 2017-02-01 西安电子科技大学 Relative ratio-based digital watermarking method
CN106373077B (en) * 2016-08-19 2019-04-23 西安电子科技大学 Digital watermark method based on relative ratio
CN106373079A (en) * 2016-08-31 2017-02-01 南京师范大学 Android photo position and time digital watermark embedding and extraction method
CN108765256A (en) * 2018-07-10 2018-11-06 天津工业大学 Data waterprint embedded method based on a kind of dct transform that human vision is covered
CN112017097A (en) * 2020-08-27 2020-12-01 豪威科技(武汉)有限公司 Watermark adding method, device, terminal and computer readable storage medium
CN112017097B (en) * 2020-08-27 2024-03-08 豪威科技(武汉)有限公司 Watermark adding method, watermark adding device, terminal and computer readable storage medium

Similar Documents

Publication Publication Date Title
CN102682418B (en) Method for embedding and extracting multiple zero watermarks of digital image
CN104268823A (en) Digital watermark algorithm based on image content
CN107067363B (en) A kind of non-blind robust Spatial Domain Image Watermark method constant based on pixel order relation
CN101122995A (en) Binary image digit water mark embedding, extraction method and device
CN102024249A (en) Digital image watermarking method based on visual perception characteristics
CN103559676A (en) Method for preventing color image blind watermarking from being printed and scanned based on DCT coefficient statistical property
CN102880999A (en) Color image digital watermark method capable of resisting stamping, printing and scanning processes
CN105512999A (en) Dual-transform color image holographic watermarking method
CN101901470A (en) Image-tampering detection and recovery method based on energy-domain semi-fragile watermarking
CN104063834A (en) Watermark embedding method, watermark detecting method and device thereof
CN104766263A (en) Color image watermark embedding and detecting method based on quaternion Legendre moment correction
CN113222804B (en) Ceramic process-oriented up-sampling ceramic watermark model training method and embedding method
CN106339978A (en) Color digital image watermark embedding and extracting method based on compressed sensing
CN109544438A (en) A kind of digital watermark method based on neural network and dct transform
Hu et al. A histogram based watermarking algorithm robust to geometric distortions
CN100421123C (en) Robust two-value text image multi-watermark realizing method
CN102156956B (en) High robustness watermark method based on singular value decomposition and discrete cosine transform
Radouane et al. A Method of LSB substitution based on image blocks and maximum entropy
CN101853483B (en) Printing and scanning resilient watermarking method based on mathematical modeling
CN101398928B (en) Digital watermarking embedding and extracting device and method for printing product
Kekre et al. Robust Watermaking Technique Using Hybrid Wavelet Transform Generated From Kekre Transforms and Discrete Cosine Transform
Zhang et al. A semi-fragile digital watermarking algorithm for 2D vector graphics tamper localization
CN102708535B (en) Zero-watermark insertion and extraction methods with multiple keys for digital images
CN103886541B (en) Image watermark method based on Tchebichef square affine invarients
CN103065073B (en) A kind of method for anti-counterfeit of the digital printed product based on digital watermark technology

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
C02 Deemed withdrawal of patent application after publication (patent law 2001)
WD01 Invention patent application deemed withdrawn after publication

Application publication date: 20140205