CN1440196A - Digital watermark implanting and extracting method based on parth search - Google Patents

Digital watermark implanting and extracting method based on parth search Download PDF

Info

Publication number
CN1440196A
CN1440196A CN 03116159 CN03116159A CN1440196A CN 1440196 A CN1440196 A CN 1440196A CN 03116159 CN03116159 CN 03116159 CN 03116159 A CN03116159 A CN 03116159A CN 1440196 A CN1440196 A CN 1440196A
Authority
CN
China
Prior art keywords
watermark
image
scramble
digital
embedding
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.)
Granted
Application number
CN 03116159
Other languages
Chinese (zh)
Other versions
CN1177472C (en
Inventor
何晨
冯国瑞
蒋铃鸽
王东建
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.)
Shanghai Jiaotong University
Original Assignee
Shanghai Jiaotong 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 Shanghai Jiaotong University filed Critical Shanghai Jiaotong University
Priority to CNB031161596A priority Critical patent/CN1177472C/en
Publication of CN1440196A publication Critical patent/CN1440196A/en
Application granted granted Critical
Publication of CN1177472C publication Critical patent/CN1177472C/en
Anticipated expiration legal-status Critical
Expired - Fee Related legal-status Critical Current

Links

Landscapes

  • Editing Of Facsimile Originals (AREA)
  • Image Processing (AREA)

Abstract

A method for embedding and extracting the digital water mark on the basis of route searching embeds the bits of the information of the digital water after disturbing the original image of the carrier according to the Chaos projection. When embedding the digital water mark, the method divides the disturbed original image into blocks, finds the gray level of the pixels of the reference point of each block after modifying the special orientation according to the bits of the water mark, making the relationship of the gray level between the pixels meet the predefined distance criterion and inverse-converts the disturbed image to obtain the image in which the water mark is embedded. When extracting the water mark, using the same disturbing and dividing manners as embedding, the method searches the route which meets the criterion

Description

Digital watermarking embedding and extracting method based on route searching
Technical field:
The present invention relates to a kind of digital watermarking embedding and extracting method based on route searching, is the copyright authentication that is applicable to the digital picture product, prevents the digital watermark method that illegally copies.
Background technology:
Digital watermark technology is by certain algorithm some significant information directly to be embedded in the middle of the content of multimedia, but does not influence the value and the use of raw content, and can not be perceiveed by people's consciousness system or notice.Watermark information can be the author sequence number, corporate logo, the text of Special Significance etc. is arranged, can be used to discern the source, version, original author, owner, publisher, legal end user of file, image or musical-product ownership to digital product.Different with encryption technology, digital watermark technology can not stop pirate movable generation, but whether it can differentiate object and be protected, and monitors that propagation, the true and false of protected data differentiated and illegally copied, solves dispute over copyright and produce evidence for court.In order to increase the difficulty of removing watermark to the assailant, present most of watermark production program all adopts encryption (comprising public-key cryptography, the private cipher key) system in the cryptography to strengthen, in the embedding of watermark, adopt a kind of key when extracting, even several key is united use.
The design of digital watermarking will be considered the following aspects:
Invisibility:, or, just be enough to show its ownership with visible sign with physical form canned data (as newpapers and periodicals, magazine) for information (as TV programme) with analog form storage and distribution.But under digital form, flag information very easily is modified or wipes.Therefore should be according to the type and the geometrical property of multimedia messages, the key that utilizes the user to provide is hidden into watermark in a series of positions that produce at random, and the people can't be discovered.
Robustness: watermark must be attacked general signal processing operations (as making an uproar, shear, strengthen, have distortion compression etc.), deletion, fascination is attacked etc. has robustness.Unless digital watermarking is had enough prioris, any destruction and the attempt of eliminating watermark are all with the quality of heavy damage multimedia messages.
Fail safe: even watermarking algorithm is open, it can prevent that equally unauthorized group from removing watermark or adding the watermark of a vacation.
The ability of hiding: it is meant can add the amount of information of watermark under the precondition that does not influence picture quality.Be used as lawful evidence in order to add enough copyright informations, watermarking algorithm should have the ability of reasonably hiding.
Watermark capacity: the watermark information of embedding must be enough to represent the founder or the possessory flag information of content of multimedia, or buyer's sequence number.When dispute over copyright took place, founder or possessory information were used for the copyright owner of unlabeled data like this, and sequence number is used to indicate and violates the agreement and the user of multi-medium data is provided for piracy.
In the existing digital watermark method, least significant bit algorithm Schyndel algorithm is considered to a kind of important watermarking algorithm, it at first produces watermark signal to m-sequencer of a key input, this m-sequence is rearranged into 2 dimension watermark signals then, and is inserted into the lowest order of original image gray value one by one by picture element.Because watermark signal has been arranged on the lowest order, it is sightless, and based on same reason, it can be removed easily, also is not strong therefore; Though it has bigger Information hiding amount, as digital watermarking algorithm, this algorithm is because of its basic principle restriction, and the digital watermark information of being hidden is very fragile, can't stand some harmless and diminish signal processing.
The Patchwork algorithm is a kind of based on the watermarking algorithm that changes the view data statistical property, this algorithm at first picked at random N to picture element, pass through to increase the brightness value of a point of pixel centering then, and the adjustment of the brightness value of another point of corresponding reduction hides Info.For increasing the robustness of its watermark, can be also the pixel area (as 8x8) of pixel to expanding to fritter, the adjustment of the brightness value of all picture elements hides Info in the corresponding minimizing corresponding region by the brightness value that increases all picture elements in the zone.But the hiding ability of this algorithm is lower.
Jui-Cheng Yen is at " Watermarks embedded in permuted domain. " (ElectronicsLetters, Volume:37, pp.80-81,2001) in a kind of watermarking algorithm has been described, and the specific implementation step, but it has a lot of limitation.At first, its image scrambling method is not fine, through its disorder method, watermark information is not distributed to entire image, but major part still concentrates on the upper left corner of image, if therefore this part is carried out shearing manipulation, the watermark information major part all can lose so, and watermark information can't propose.Moreover the ability of its anti-JPEG (Joint Photographic Experts Group) compression standard neither be fine, and under 40% compression ratio, the watermark information of extraction can't be confirmed substantially.At last, its noise resisting ability is not strong, and after image added small amount of noise, watermark information can't be recognized.
Summary of the invention:
The objective of the invention is at the deficiencies in the prior art, a kind of digital watermarking embedding and extracting method based on route searching proposed, simple, the attack that carrier image suffered is had good resistivity, after the image that adds watermark is under attack, still can recover original watermark substantially.
For realizing such purpose, method of the present invention at first utilizes the character of chaotic maps that original image is carried out scramble, in spatial domain, watermark information is embedded in the gray-scale relation between pixel then,, revises the gray-scale relation of different pixels collection according to the difference of bit information.In order to obtain watermark bit, be equivalent to seek respective pixel in the set of pixels during extraction, judge the relation between pixel,, just can extract watermark bit as long as they satisfy gray-scale relation criterion when embedding.Watermark embeds to make satisfies certain distance criterion between pixel grey scale, it can be the ultimate range criterion, and minimum distance criterion, or a self-defining criterion will be with same setting criterions during extraction.In fact it be equivalent to have between per two pixels in the set of pixels one to connect arc, gives different weights on every arc, and information bit sequence is mapped as a paths, revises the weights on all arcs of this paths, makes weights satisfy desired distance criterion.The path method that distance criterion is satisfied in search has a variety of, comprises traversal search, Viterbi algorithm search, sequential decoding algorithm search, stack algorithm search or the like.Equally, also can utilize this method that watermark bit is embedded in the transform domain, as Fourier transform, cosine transform domain, wavelet transformed domain.In contrast to present algorithm, it has lot of advantages, the JPEG compression, and making an uproar of image all has good resistivity.
Concrete steps of the present invention comprise:
1, according to chaotic maps the initial carrier image scrambling, adoptable method is a lot of in this step, for example directly with chaotic maps entire image is carried out whole scramble, perhaps the random point that produces with chaos sequence can reach the scramble purpose to the mode of mutual exchange;
2, embed digital watermark information bit in the image behind scramble.When digital watermarking embeds, the original image piecemeal behind the scramble, find every datum mark to revise the grey scale pixel value of particular orientation according to watermark bit, make gray-scale relation between pixel satisfy the distance criterion of certain predefined, then image is done the scramble inverse transformation, obtain watermarked image;
Watermark embeds to make satisfies certain distance criterion between the pixel grey scale, it can be the ultimate range criterion, minimum distance criterion, or a self-defining criterion.Again image is carried out the scramble inverse transformation at last, could not influence original image information like this, this moment, telescopiny finished.
3, during watermark extracting, at first the watermarking images process disorder method effect identical of receiving with embedding, and then piecemeal, in each piece, find out datum mark, satisfy the path that embeds distance criterion from one of datum mark search, determine the pixel of this paths process, recover the corresponding digital watermark bit according to their orientation;
To use same setting criterion when embedding during watermark extracting with digital watermarking.In fact it be equivalent to have between per two pixels in the set of pixels one to connect arc, gives different weights on every arc, and information bit sequence is mapped as a paths, revises the weights on all arcs of this paths, makes weights satisfy desired distance criterion.The path method that distance criterion is satisfied in search has a variety of, comprises traversal search, Viterbi algorithm search, sequential decoding algorithm search, stack algorithm search or the like.Equally, also can utilize this method that watermark bit is embedded in the transform domain, as Fourier transform, cosine transform domain, wavelet transformed domain.
4, utilize the watermark bit of extracting, form watermarking images.
The present invention utilizes the character of chaotic maps that original image is carried out scramble, experimental results show that its effect is fine, and complexity is not high, compares with traditional disorder method, and is simple and easily understand, and speed is very fast when adopting overall scramble.When digital watermarking embeds, watermark information is embedded in the relation between pixel, can not extract any watermark bit from single pixel, can only extract watermark bit by the relation of compared pixels, during extraction, require the whole piece path to satisfy the optimal distance of predefined, so also it doesn't matter when the gray scale of some point changed on this paths, has certain error correcting capability.
Method of the present invention is simple, and original image is not degraded basically; It has good resistivity to the common attack that carrier image suffered, especially for shearing, and Gaussian noise, the JPEG compression after the image that adds watermark is under attack, still can recover original watermark substantially.
The new image watermark that the present invention proposes embeds and extracting method, is specially adapted to the copyright authentication of digital picture product, prevents the digital watermark method that illegally copies, protection copyright owner's interests and their intellectual property.
Description of drawings:
Fig. 1 is digital watermarking embedding of the present invention, extraction algorithm FB(flow block).
Original " Lena " of Fig. 2 (a) figure, (b) watermarking images, " Lena " figure after (c) watermarked;
Fig. 3 (a) and (b) be after JPEG is compressed to 55% and 36%, the watermarking images of extraction;
Fig. 4 (a) adds PSNR=40, PSNR=37, and after the Gaussian noise of PSNR=34, the watermarking images of extraction (PSNR is exactly a Y-PSNR).
Embodiment:
Below in conjunction with accompanying drawing technical scheme of the present invention is further described.
Method of the present invention mainly comprises image scrambling, adds watermark and extracts watermark, as shown in Figure 1.
In the image scrambling process, at first adopt a chaos sequence that original image is carried out whole scramble, scramble need be controlled by original image authorized person's key (1), if do not obtain this key, any entity or individual can't repeat this scramble process.
Adding the watermark process, at first the original image piecemeal behind the scramble, find out the coordinate at every center, select one of them piece, with the center is datum mark, revise the consecutive points of different directions according to the difference of encrypting the bit sequence value, revise the distance criterion of rule according to prior agreement, for example, can define such criterion, be 1 if encrypt bit sequence, then revise the right-hand adjacent point of datum mark, making between it and the datum mark and satisfy apart from minimum criteria, is 0 if encrypt bit sequence, then revises an adjacent point of datum mark left, make and satisfy between it and the datum mark apart from minimum criteria, next step is 1 if encrypt bit sequence, revises a point of datum mark top, make the distance between the point of it and last change satisfy minimum criteria, if ciphering sequence is 0, revise a point of datum mark below, make the distance between the point of it and last change satisfy minimum criteria.With distance between reference be that 1 point has been revised, revise distance again and be 2 point, identical with said method, arrive up to the point of revising till the border of piecemeal, at this moment select other piece to add watermark in the same way.
Adding the encrypted watermark image too need be by original image authorized person's key (2) control.
Add in the image transmission course of encrypted watermark, may be subjected to illegal authorized person's attack.When image received, identical method was carried out the image scrambling inverse transformation when needing to adopt with the scramble original image.
Corresponding to watermark extraction process, as long as the central point of corresponding each piece is taken out, the path of distance minimum is found out, concern according to the position of itself and central point, just can rule out watermark.Utilize identical method, also can once add dibit or three bit encryption watermarks, the former only needs to consider the four direction adjacent with central point, and the latter considers that eight adjacent directions get final product.At last again image behind the scramble is returned to original form, so just watermark information is spread to the various piece of image.
In extracting the watermark process, adopt and the identical scramble mode of interpolation watermark process, find every datum mark simultaneously, according to watermarked method, find a series of pixels that satisfy distance criterion, judge direction, find the embedding bit.
Embodiment:
If original image is standard " Lena " (Fig. 2 .a), size is 512 * 512, and watermark is the image (Fig. 2 .b) that 64 * 64 two-value Chinese " leads to " word.
At first image is carried out overall scramble, adopts following Logistic mapping:
x n+1=μx n(1-x n) 0<x n<1
When μ=3.9, and appropriately choose initial value x 0=0.8 (key), the ordered series of numbers x of generation nHas chaotic characteristic.Give up preceding 9 values, since the 10th value, each value binary expansion, get preceding 36 bits, per 9 bits are formed an integer, form 4 integers like this, the row of preceding two integer correspondences is exchanged, and the row of latter two integer correspondence exchange, and so repeat 1000 times.Two-dimentional watermark figure is become a bit stream b (i).Again image is divided into 16 * 16 block of pixels, the central point of establishing every is that the upper left corner begins to the right and moves down 7 pixels, is designated as (p j, q j) adopt foregoing implementation process, add the encrypted watermark bit from first BOB(beginning of block), specific algorithm is as follows:
Get first as an example, suppose d Xy(x, y) (mod32) requires 0≤d here to=f Xy<32, definition floor (f (x y), c) is:
Floor (f (x, y), c)=f (x, y)+c-d XyA) gray value of modification central point:
f New(p 1, q 1)=floor (f (p 1, q 1), 24) b) revise grey scale pixel value along horizontal ordinate successively by following rule:
When i is even number, if b (i)=1, f new ( p 1 + 1 + i 2 , q 1 ) = floor ( f ( p 1 + 1 + i 2 , q 1 ) , 24 ) f new ( p 1 + 1 + i 2 , q 1 ) = floor ( f ( p 1 + 1 + i 2 , q 1 ) , 16 )
Otherwise f new ( p 1 + 1 + i 2 , q 1 ) = floor ( f ( p 1 + 1 + i 2 , q 1 ) , 24 ) f new ( p 1 + 1 + i 2 , q 1 ) = floor ( f ( p 1 + 1 + i 2 , q 1 ) , 16 )
I=i+1.c then) when i is odd number, if b (i)=1, f new ( p 1 , q 1 + i 2 ) = floor ( f ( p 1 , q 1 + i 2 ) , 24 ) f new ( p 1 , q 1 - i 2 ) = floor ( f ( p 1 , q 1 - i 2 , q 1 ) , 0 )
Otherwise f new ( p 1 , q 1 - i 2 ) = floor ( f ( p 1 , q 1 - i 2 ) , 24 ) f new ( p 1 , q 1 + i 2 ) = floor ( f ( p 1 , q 1 + i 2 ) , 0 )
I=i+1. then
Each step above repeating is till the border that is modified to piece.Choose other piece successively and also add the residue watermark bit in the same way.At last image is carried out the scramble inverse transformation, form Fig. 2 .c.
In the leaching process, identical disorder method and partitioned mode when adopting with embedding, find datum mark then, judge the each point of four direction, obtain an optimal path according to minimum distance criterion, on this path each put the just corresponding watermark bit of pairing direction, so repeats up to having extracted whole watermark bit.
Through being compressed to after 55% and 36% JPEG attacks, the watermarking images that extracts, see Fig. 3 .a and 3.b, at process PSNR=40, PSNR=37, after the PSNR=34 Gaussian noise is attacked, the watermarking images of extraction, be respectively Fig. 4 .a, 4.b and 4.c, therefrom we still can pick out the information that watermark comprises.

Claims (1)

1. digital watermarking embedding and extracting method based on a route searching, its feature comprises the steps:
1) according to chaotic maps the initial carrier image scrambling;
2) embed digital watermark information bit in the image behind scramble, when digital watermarking embeds, the original image piecemeal behind the scramble, find every datum mark to revise the grey scale pixel value of particular orientation according to watermark bit, make gray-scale relation between pixel satisfy the distance criterion of certain predefined, then image is done the scramble inverse transformation, obtain watermarked image;
3) during watermark extracting, at first adopt the scramble mode identical to be handled with the embedding stage to the image of receiving, and then piecemeal, in each piece, find out datum mark, satisfy the path that embeds distance criterion from one of datum mark search, determine the pixel of this paths process, recover the corresponding digital watermark bit according to their orientation;
4) utilize the watermark bit of extracting, form watermarking images.
CNB031161596A 2003-04-03 2003-04-03 Digital watermark implanting and extracting method based on parth search Expired - Fee Related CN1177472C (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CNB031161596A CN1177472C (en) 2003-04-03 2003-04-03 Digital watermark implanting and extracting method based on parth search

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CNB031161596A CN1177472C (en) 2003-04-03 2003-04-03 Digital watermark implanting and extracting method based on parth search

Publications (2)

Publication Number Publication Date
CN1440196A true CN1440196A (en) 2003-09-03
CN1177472C CN1177472C (en) 2004-11-24

Family

ID=27797095

Family Applications (1)

Application Number Title Priority Date Filing Date
CNB031161596A Expired - Fee Related CN1177472C (en) 2003-04-03 2003-04-03 Digital watermark implanting and extracting method based on parth search

Country Status (1)

Country Link
CN (1) CN1177472C (en)

Cited By (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101286224B (en) * 2008-05-13 2010-06-02 南京航空航天大学 Watermark embedding and extraction method based on fractal image coding
CN101241543B (en) * 2008-02-28 2011-03-30 天津市阿波罗信息技术有限公司 Processing method for embedding information at commodity packaging recognized by mobile phone
CN102123327A (en) * 2010-12-23 2011-07-13 上海交通大学 Method for embedding and extracting digital watermark on basis of streaming media noncritical frame
CN102142129A (en) * 2011-03-09 2011-08-03 西安理工大学 Visual secret sharing-based image content tampering detection method
CN101389027B (en) * 2008-09-28 2012-08-08 华为技术有限公司 Watermark embedding method, watermark extracting method, device and system
CN102646260A (en) * 2012-02-17 2012-08-22 中山大学 Robustness digital watermark method based on chaotic mapping and singular value decomposition
CN111915472A (en) * 2019-05-10 2020-11-10 浙江大学 Watermark embedding method, watermark extracting method and watermark extracting device

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101877119B (en) * 2009-11-20 2012-01-11 南京师范大学 Data format adaptive method for embedding and extracting GIS vector data watermark

Cited By (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101241543B (en) * 2008-02-28 2011-03-30 天津市阿波罗信息技术有限公司 Processing method for embedding information at commodity packaging recognized by mobile phone
CN101286224B (en) * 2008-05-13 2010-06-02 南京航空航天大学 Watermark embedding and extraction method based on fractal image coding
CN101389027B (en) * 2008-09-28 2012-08-08 华为技术有限公司 Watermark embedding method, watermark extracting method, device and system
CN102123327A (en) * 2010-12-23 2011-07-13 上海交通大学 Method for embedding and extracting digital watermark on basis of streaming media noncritical frame
CN102123327B (en) * 2010-12-23 2012-12-26 上海交通大学 Method for embedding and extracting digital watermark on basis of streaming media noncritical frame
CN102142129A (en) * 2011-03-09 2011-08-03 西安理工大学 Visual secret sharing-based image content tampering detection method
CN102142129B (en) * 2011-03-09 2012-07-04 西安理工大学 Visual secret sharing-based image content tampering detection method
CN102646260A (en) * 2012-02-17 2012-08-22 中山大学 Robustness digital watermark method based on chaotic mapping and singular value decomposition
CN111915472A (en) * 2019-05-10 2020-11-10 浙江大学 Watermark embedding method, watermark extracting method and watermark extracting device
CN111915472B (en) * 2019-05-10 2023-07-28 浙江大学 Watermark embedding method, watermark extracting method and watermark embedding device

Also Published As

Publication number Publication date
CN1177472C (en) 2004-11-24

Similar Documents

Publication Publication Date Title
Chao et al. A novel image data hiding scheme with diamond encoding
Hu et al. Reversible visible watermarking and lossless recovery of original images
Zhu et al. A new semi-fragile image watermarking with robust tampering restoration using irregular sampling
CN109859093B (en) Mixed transform domain image zero-watermarking method based on variable parameter chaotic mapping
CN1177472C (en) Digital watermark implanting and extracting method based on parth search
Arya et al. Improved capacity image steganography algorithm using 16-pixel differencing with n-bit LSB substitution for RGB images
CN115311119A (en) Three-dimensional image zero watermark embedding and extracting method capable of resisting geometric attack
CN1195378C (en) Chaos based robustness invisuable implanting and extracting method
Li et al. Oblivious fragile watermarking scheme for image authentication
Surekha et al. Visual secret sharing based digital image watermarking
Surekha et al. A multiple watermarking technique for images based on visual cryptography
Gao et al. Reversible watermarking algorithm based on wavelet lifting scheme
Wu et al. A chaos based robust spatial domain watermarking algorithm
Saryazdi et al. A blind digital watermark in Hadamard domain
Surekha et al. A watermarking technique based on visual cryptography
Zhang et al. Blind dual watermarking scheme using stucki kernel and spiht for image self-recovery
CN112862654B (en) Blind watermark embedding and extracting method based on combination of relational interpolation and linear interpolation
Meryem et al. A short survey on image zero-watermarking techniques based on visual cryptography
Cauvery Water marking on digital image using genetic algorithm
Zhong et al. A spatial domain color watermarking scheme based on chaos
Punitha et al. Notice of Violation of IEEE Publication Principles: Binary watermarking technique based on Visual Cryptography
Tsai et al. A Steganographic Method by Pixel–Value Differencing and Exploiting Modification Direction
Surekha et al. Digital image ownership verification based on spatial correlation of colors
Bassali et al. Compression tolerant watermarking for image verification
Sumalatha et al. Local content based image authentication for tamper localization

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
C14 Grant of patent or utility model
GR01 Patent grant
C19 Lapse of patent right due to non-payment of the annual fee
CF01 Termination of patent right due to non-payment of annual fee