CN108648130A - A kind of total blindness's digital watermark method with copyright protection and tampering location function - Google Patents
A kind of total blindness's digital watermark method with copyright protection and tampering location function Download PDFInfo
- Publication number
- CN108648130A CN108648130A CN201810299604.9A CN201810299604A CN108648130A CN 108648130 A CN108648130 A CN 108648130A CN 201810299604 A CN201810299604 A CN 201810299604A CN 108648130 A CN108648130 A CN 108648130A
- Authority
- CN
- China
- Prior art keywords
- region
- denoted
- block
- watermark
- sub
- 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
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T1/00—General purpose image data processing
- G06T1/0021—Image watermarking
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2201/00—General purpose image data processing
- G06T2201/005—Image watermarking
- G06T2201/0065—Extraction of an embedded watermark; Reliable detection
Abstract
The invention discloses a kind of total blindness's digital watermark methods with copyright protection and tampering location function, and nonoverlapping 8 × 8 sub-block is divided into the ll channel of original-gray image, each sub-block is divided into 4 × 4 region；Discrete cosine transform is carried out to each region；Feature watermark and self-embedding feature watermark are created according to the dct coefficient matrix of different zones；Discrete cosine transform is carried out to each of ll channel of watermarking images each region in the block in an identical manner when extraction；Feature watermark and authenticating water-mark are blindly extracted then according to the dct coefficient matrix of different zones；It calculates the normalizated correlation coefficient between feature watermark and authenticating water-mark and carries out copyright protection, successive appraximation feature watermark and authenticating water-mark realize total blindness's tampering location；Advantage is only need to be embedded in that a digital watermarking can be achieved with copyright protection and tampering location is difunctional, any information when extracting watermark without original image and original figure watermark, can realize total blindness's detection of watermark.
Description
Technical field
The present invention relates to a kind of digital media information safe practice, there is copyright protection and distort more particularly, to one kind and determine
Total blindness's digital watermark method of bit function.
Background technology
With the rapid development of Digital Media and network technology, digital watermark technology has become medium copyright protection and content
A kind of effective means of integrated authentication.But existing digital watermark technology is mostly single watermarking algorithm, often there is functions
Single limitation can only realize the protection of copyright or can only realize the certification of content.Obviously, there is multi-functional digital water
Print algorithm has stronger demand in practical applications.
Multifunction digital watermark refers to the embedded watermark of different nature in same width digital picture, is answered with reaching different
Use purpose.Such as：It is embedded in a visible watermark and an invisible watermark simultaneously in a width digital picture, the former is used for copyright
Notice, the latter are used for copyright protection.For another example：It is embedded in a fragile watermark and a robust water simultaneously in a width digital picture
Print, the former is for distorting prompt and content authentication, and the latter is for copyright protection etc..
2009, Ye Tianyu existed《Electronics and information journal》On propose a kind of multi-functional double watermarks algorithm, which exists
Embedded robust watermarking in the singular value of image block, it is embedded fragile on the spatial domain pixel least significant bit of the image containing robust watermarking
Watermark realizes copyright protection and content authentication dual function with this.2012, Lin Junhai existed《Electronic University Of Science ＆ Technology Of Hangzhou's journal》
On propose a kind of multifunctional watermark algorithm that null tone domain based on Error Correction of Coding combines, which carries out carrier image discrete
Multipurpose watermarking is embedded into the diagonal line subband of low frequency sub-band by wavelet transformation, then utilizes the bit plane and robust of image
Property watermark carry out Hamming code coding, the check code of generation is embedded in the least significant bit of image, realizes the insertion of fragile watermark.
2013, Yu Xiaoqing existed《Computer and application software》On propose a kind of double water of the Multifunctional color color image based on Lifting Wavelet
Algorithm is printed, which utilizes the different characteristics of three Color Channels of color RGB image, by the low frequency coefficient of blue component
Embedded robust watermarking, the embedded semi-fragile watermarking on the intermediate frequency coefficient of green component, realizes image copyright protection and content authentication
And the multi-functional of tampering location.2015, Wang Juan《Minnan Normal University's journal》On propose a kind of more work(of coloured image
Can double watermarks algorithm, the algorithm by extract coloured image tri- channels RGB, by copyright watermark qualified insertion in blue component
Discrete wavelet low frequency coefficient piecemeal singular value in, then the authenticating water-mark adaptive quantizing generated from red component is embedded in
In green component piecemeal singular value, the experimental results showed that, which has the dual function of copyright protection and content authentication.
As it can be seen that in order to realize the multi-functional purpose of digital watermarking, it is set forth above that there is multi-functional digital watermarking algorithm
It needs robust watermarking and fragile watermark (or authenticating water-mark) while being embedded in image-carrier.Obviously, it is calculated compared to single watermark
Method, while doubly time series model insertion inevitably lead to the decline of picture quality；And due to being carried in same image
Embedded doubly time series model in body, therefore certain interference is certainly existed between robust watermarking and fragile watermark, after influencing
The extraction effect of continuous watermark.In addition, in order to realize the functions such as copyright protection and content authentication, to the robust watermarking and fragility of extraction
Watermark needs and original figure watermark carry out respectively relativity measurement and compared with, cannot achieve the blind Detecting of digital watermarking.
Therefore, for overcome it is above have multi-functional digital watermarking algorithm existing for associated limitations, research a kind of it is embedding
Enter single watermark and can be achieved with multi-functional purpose, and can realize that the digital watermarking algorithm of total blindness's detection more has significant practical applications,
Also will be the developing direction of field of information security technology.
Invention content
Technical problem to be solved by the invention is to provide a kind of total blindness's numbers with copyright protection and tampering location function
Word water mark method only needs an embedded digital watermarking to can be achieved with double work(of copyright protection and tampering location in built-in end
Can, and under the premise of ensureing watermarking images quality, there is more satisfactory robustness for normal image processing, for malice
Attack has good sensibility, any letter when digital watermarking is extracted in test side without original image and original figure watermark
Breath can realize total blindness's detection of digital watermarking.
Technical solution is used by the present invention solves above-mentioned technical problem：One kind having copyright protection and tampering location work(
Can total blindness's digital watermark method, it is characterised in that main processes are：
At digital watermark embedding end, level-one discrete wavelet transformation is carried out to original 8bit gray level images, is forced what is obtained
The sub-block that nearly dividing sub-picture is 8 × 8 at the size of multiple non-overlapping copies, 4 are divided by each sub-block in ll channel
The region that the size of a non-overlapping copies is 4 × 4；Then to each of ll channel each region in the block carry out from
Dissipate cosine transform；Then according to the discrete cosine transform of the upper left corner area and upper right comer region of all sub-blocks in ll channel
DC coefficient, establishment obtain feature watermark；Lower left corner region further according to each sub-block in ll channel and lower right field
Same position discrete cosine transform ac coefficient and create every bit in obtained feature watermark, in ll channel
Each sub-block in a bit in self-embedding feature watermark；Finally by inverse discrete cosine transform and level-one against discrete wavelet
Transformation obtains the watermarking images embedded with digital watermarking；
In digital watermarking extraction and test side, level-one discrete wavelet transformation is carried out to the watermarking images embedded with digital watermarking,
Obtained ll channel is divided into the sub-block that the size of multiple non-overlapping copies is 8 × 8, by every height in ll channel
Block is divided into the region that the size of 4 non-overlapping copies is 4 × 4；Then in the block to each of ll channel each
Region carries out discrete cosine transform；Upper left corner area and upper right comer region then according to all sub-blocks in ll channel from
Cosine transform DC coefficient is dissipated, Blind extracting obtains feature watermark；Further according to the lower left corner region of each sub-block in ll channel
With the discrete cosine transform ac coefficient of the same position of lower right field, Blind extracting obtains authenticating water-mark；Finally by calculating
Normalizated correlation coefficient between the authenticating water-mark that the feature watermark and Blind extracting that Blind extracting obtains obtain carries out copyright protection,
Authenticating water-mark that the feature watermark and Blind extracting obtained by successive appraximation Blind extracting obtains realizes total blindness's tampering location.
Total blindness's digital watermark method with copyright protection and tampering location function includes digital watermark embedding, digital water
Two parts of print extraction and detection；
The digital watermark embedding part the specific steps are：
Step 1. _ 1, choose an original 8bit gray level image, and be denoted as F, the size of F is I × J；Then to F
Level-one discrete wavelet transformation is carried out, the three width details subgraphs and a width ll channel of F is obtained, the ll channel of F is denoted as FA, FA
Size beThen FA is divided intoThe sub-block that the size of a non-overlapping copies is 8 × 8, will
K-th of sub-block in FA is denoted as FABk；It is 4 × 4 that each sub-block in FA, which is divided into the size of 4 non-overlapping copies, again
Region, by FABkIn upper left corner area as FABkIn the 1st region be denoted asBy FABkIn upper right comer region make
For FABkIn the 2nd region be denoted asBy FABkIn lower left corner region as FABkIn the 3rd region be denoted asBy FABkIn lower right field as FABkIn the 4th region be denoted asWherein, I indicates the level side of F
To resolution ratio, J indicates the vertical direction resolution ratio of F, symbolFor downward rounding operation symbol, k is positive integer,
1. _ 2, to each of FA each region in the block step carries out discrete cosine transform, obtain each of FA
The dct coefficient matrix in son each region in the block, willDct coefficient matrix be denoted asIt willDct coefficient matrix be denoted asIt willDct coefficient matrix note
ForIt willDct coefficient matrix be denoted asWherein, WithDimension be 4 × 4；
Step 1. _ 3, create feature watermark W, detailed process is：The discrete cosine transform for calculating each sub-block in FA is straight
Coefficient Mean is flowed, by FABkDiscrete cosine transform DC coefficient mean value be denoted as FADk,
Then according to the discrete cosine transform DC coefficient mean value of all sub-blocks in FA, establishment obtains feature watermark W, by the kth position of W
Bit is denoted as Wk,Wherein, the length of W is It indicatesIn
The 1st column element of the 1st row,ForDiscrete cosine transform DC coefficient,It indicatesIn
The 1st column element of 1st row,ForDiscrete cosine transform DC coefficient, δ1For the first threshold of setting；
Step 1. _ 4, self-embedding feature watermark W, detailed process is：Self-embedding feature watermark W in each sub-block in FA
In a bit, for FABk, take out WkIf Wk=0 andThen enableThen it enablesRealize Wk's
Self-embedding, then whenIt is right when establishmentWithMake
Further adjustment, evenAnd it enables
If Wk=0 andIt is then rightWithIt does not deal with, it is real
Existing WkSelf-embedding, then whenIt is right when establishmentWithIt is for further adjustments, evenAnd it enables
If Wk=1 andThen enableThen it enablesRealize WkSelf-embedding, then when
It is right when establishmentWithIt is for further adjustments, evenAnd it enables
If Wk=1 andIt is then rightWithIt does not deal with, it is real
Existing WkSelf-embedding, then whenIt is right when establishmentWithIt is for further adjustments, evenAnd it enables
Wherein,It indicatesIn the 2nd column element of the 2nd row,It indicatesIn the 2nd row
2nd column element, G1 and G2 are the intermediate variable of introducing, and abs () is the function that takes absolute value, δ2For the second threshold of setting, In "=" be assignment；
Step 1. _ 5, step 1. _ 4 on the basis of, to each region carry out inverse discrete cosine transform；Then recombination obtains
Each sub-block for being embedded with digital watermarking, and then recombinate and obtain the ll channel embedded with digital watermarking；Then to being embedded with digital watermarking
Ll channel and the three width details subgraphs of F carry out level-one inverse discrete wavelet transform, obtain the watermarking images embedded with digital watermarking；
The digital watermarking extraction and detection part the specific steps are：
Step 2. _ 1, the watermarking images embedded with digital watermarking be 8bit gray level images, be denoted as TF, the size of TF is I
×J；Then level-one discrete wavelet transformation is carried out to TF, the three width details subgraphs and a width ll channel of TF is obtained, by forcing for TF
Nearly subgraph is denoted as TFA, and the size of TFA isThen TFA is divided intoThe size of a non-overlapping copies is big
K-th of sub-block in TFA is denoted as TFAB by the small sub-block for being 8 × 8k；It is divided into 4 not weigh mutually each sub-block in TFA again
The region that folded size is 4 × 4, by TFABkIn upper left corner area as TFABkIn the 1st region be denoted asBy TFABkIn upper right comer region as TFABkIn the 2nd region be denoted asBy TFABkIn lower-left
Angular zone is as TFABkIn the 3rd region be denoted asBy TFABkIn lower right field as TFABkIn the 4th
Region is denoted asWherein, symbolFor downward rounding operation symbol, k is positive integer,
2. _ 2, to each of TFA each region in the block step carries out discrete cosine transform, obtain every in TFA
The dct coefficient matrix in height each region in the block, willDct coefficient matrix be denoted asIt willDct coefficient matrix be denoted asIt willDiscrete cosine transform coefficient square
Battle array is denoted asIt willDct coefficient matrix be denoted asWherein, WithDimension be 4 × 4；
Step 2. _ 3, Blind extracting feature watermark TW, detailed process is：The discrete cosine for calculating each sub-block in TFA becomes
DC coefficient mean value is changed, by TFABkDiscrete cosine transform DC coefficient mean value be denoted as TFADk,Then according to the discrete cosine transform direct current system of all sub-blocks in TFA
Number mean value, Blind extracting obtain feature watermark TW, the kth position bit of TW are denoted as TWk,Its
In, the length of TW is It indicatesIn the 1st column element of the 1st row,ForDiscrete cosine transform DC coefficient,It indicatesIn the 1st column element of the 1st row,
ForDiscrete cosine transform DC coefficient, δ1For the first threshold of setting；
Step 2. _ 4, Blind extracting authenticating water-mark TW*, detailed process is：One is blindly extracted in each sub-block in TFA
Bit is to obtain authenticating water-mark TW*, for TFABk, extract and obtain TW*In kth position bit TW* k,Wherein, TW*Length be Table
ShowIn the 2nd column element of the 2nd row,It indicatesIn the 2nd column element of the 2nd row；
Step 2. _ 5, according to TW and TW*Copyright protection is carried out, detailed process is：Calculate TW and TW*Between normalization phase
Relationship number is denoted as ρ (TW, TW*), utilize ρ (TW, TW*) carry out copyright protection；
According to TW and TW*Tampering location is carried out, detailed process is：According to the every bit and TW in TW*In corresponding ratio
Spy, determines whether each sub-block in TFA is tampered with, for TFABkIf TWkWith TW* kIt is equal, it is determined that TFABkIt is not usurped
It corrects one's mistakes；If TWkWith TW* kIt is unequal, it is determined that TFABkIt is tampered with, realizes total blindness's tampering location.
The step 2. _ 5 in,Its
In,Indicate the average value of all elements in TW,Indicate TW*In all elements average value.
Compared with the prior art, the advantages of the present invention are as follows：
1) the method for the present invention only needs watermarking images to can be carried out digital watermarking to carry in digital watermarking extraction with test side
It takes and digital watermark detecting, it is no longer necessary to digital watermark embedding end original image or any relevant information of original figure watermark,
Not only save digital watermark embedding end transmission original image or when original figure watermark required transmission cost and storage at
This, and the passive attack of generally existing or interpretation Attack on internet are avoided, fully achieve blind examination brake.
2) the method for the present invention digital watermark embedding end only need an embedded feature watermark achieve that copyright protection and
The difunctional purpose of tampering location, avoids the interference between more watermarks；And the establishment of feature watermark is being schemed respectively with insertion
It is carried out in the different zones of each sub-block as in, avoids the influence that feature watermark telescopiny extracts itself.
3) the method for the present invention can realize copyright protection and the difunctional purpose of tampering location simultaneously, have widely practical
Property.
4) under the premise of ensureing watermarking images quality, there is more satisfactory robustness for normal image processing, it is right
There is good sensibility in malicious attack.
Description of the drawings
Fig. 1 is the original Lena gray level images that resolution ratio is 512 × 512；
Fig. 2 is to be obtained after being handled original Lena gray level images shown in FIG. 1 using the method for the present invention embedded with water
The watermark Lena gray level images of print；
Fig. 3 is to carry out histogram equalization processing to watermark Lena gray level images shown in Fig. 2, and what is obtained is equal through histogram
Weighing apparatusization treated watermark Lena gray level images；
Fig. 4 is to carry out JPEG lossy compressions processing (compression quality selecting predictors to watermark Lena gray level images shown in Fig. 2
For the watermark Lena gray level images after JPEG lossy compressions 5%), obtained；
Fig. 5 is that (noise of superposition is that mean value is 0 to watermark Lena gray level images shown in Fig. 2 progress noise overlap-add procedure
And variance be 0.001 Gaussian reflectivity mirrors), the watermark Lena gray level images after obtained superimposed Gaussian noise；
Fig. 6 is to carry out the median filter process (window size of median filter to watermark Lena gray level images shown in Fig. 2
It is selected as [5 × 5]), the obtained watermark Lena gray level images after medium filtering；
Fig. 7 is the baboon gray level images that size is 128 × 128；
Fig. 8 is by (the 109 of watermark Lena gray level images shown in Fig. 2:236,149:276) region is substituted for shown in Fig. 7
Baboon gray level images, the watermark Lena gray level images obtained after replacement operation；
Fig. 9 is the testing result for carrying out tampering location to watermark Lena gray level images shown in Fig. 8 using the method for the present invention；
Figure 10 is by (the 129 of watermark Lena gray level images shown in Fig. 2:256,129:256) pixel in region replaces
Change another region (229 of images themselves into:356,229:356) pixel in, the watermark Lena ashes obtained after replacement operation
Spend image；
Figure 11 is the detection knot for carrying out tampering location to watermark Lena gray level images shown in Fig. 10 using the method for the present invention
Fruit；
Figure 12 is by (the 329 of watermark Lena gray level images shown in Fig. 2:456,229:356) region is substituted for shown in Fig. 1
Original Lena gray level images (229:356,379:506) region, the watermark Lena gray level images obtained after replacement operation；
Figure 13 is the detection knot for carrying out tampering location to watermark Lena gray level images shown in Figure 12 using the method for the present invention
Fruit；
Figure 14 is that watermark Lena gray level images shown in Fig. 2 are located at (214:267,189:227) image block in region
It cuts away, the watermark Lena gray level images after obtained region cutting operation；
Figure 15 is the detection knot for carrying out tampering location to watermark Lena gray level images shown in Figure 14 using the method for the present invention
Fruit；
Figure 16 is to be located at (329 to watermark Lena gray level images shown in Fig. 2:456,129:256) image block in region
Increase into row pixel value and operate, obtained pixel value increases the watermark Lena gray level images after operation；
Figure 17 is the detection knot for carrying out tampering location to watermark Lena gray level images shown in Figure 16 using the method for the present invention
Fruit；
Figure 18 is that the overall of the method for the present invention realizes block diagram.
Specific implementation mode
Below in conjunction with attached drawing embodiment, present invention is further described in detail.
A kind of total blindness's digital watermark method with copyright protection and tampering location function proposed by the present invention, it is overall real
Existing block diagram is as shown in figure 18, and main processes are：
At digital watermark embedding end, level-one discrete wavelet transformation is carried out to original 8bit gray level images, is forced what is obtained
The sub-block that nearly dividing sub-picture is 8 × 8 at the size of multiple non-overlapping copies, then each sub-block in ll channel is divided into
The region that the size of 4 non-overlapping copies is 4 × 4；Then each of ll channel each region in the block is carried out
Discrete cosine transform；The discrete cosine of upper left corner area and upper right comer region then according to all sub-blocks in ll channel becomes
DC coefficient is changed, establishment obtains feature watermark；Lower left corner region further according to each sub-block in ll channel and bottom right angular region
Every bit in the feature watermark that the discrete cosine transform ac coefficient of the same position in domain and establishment obtain, in ll channel
In each sub-block in a bit in self-embedding feature watermark；Finally by inverse discrete cosine transform and level-one against discrete small
Wave conversion obtains the watermarking images embedded with digital watermarking；
In digital watermarking extraction and test side, level-one discrete wavelet transformation is carried out to the watermarking images embedded with digital watermarking,
Obtained ll channel is divided into the sub-block that the size of multiple non-overlapping copies is 8 × 8, then by each of ll channel
Sub-block is divided into the region that the size of 4 non-overlapping copies is 4 × 4；Then in the block to each of ll channel every
A region carries out discrete cosine transform；Upper left corner area and upper right comer region then according to all sub-blocks in ll channel
Discrete cosine transform DC coefficient, Blind extracting obtain feature watermark；Further according to the lower-left angular region of each sub-block in ll channel
The discrete cosine transform ac coefficient of the same position in domain and lower right field, Blind extracting obtain authenticating water-mark；Finally by meter
The normalizated correlation coefficient between the authenticating water-mark that the obtained feature watermark of Blind extracting and Blind extracting obtain is calculated to carry out copyright guarantor
Shield, authenticating water-mark that the feature watermark and Blind extracting obtained by successive appraximation Blind extracting obtains realize total blindness's tampering location.
Total blindness's digital watermark method with copyright protection and tampering location function of the present invention include digital watermark embedding,
Digital watermarking is extracted and two parts of detection；
The digital watermark embedding part the specific steps are：
Step 1. _ 1, choose an original 8bit gray level image, and be denoted as F, the size of F is I × J；Then to F
Level-one discrete wavelet transformation is carried out, the three width details subgraphs and a width ll channel of F is obtained, the ll channel of F is denoted as FA, FA
Size beThen FA is divided by raster scan orderThe size of a non-overlapping copies
For 8 × 8 sub-block, k-th of sub-block in FA is denoted as FABk；Each sub-block in FA is divided into 4 non-overlapping copies again
The region that size is 4 × 4, by FABkIn upper left corner area as FABkIn the 1st region be denoted asBy FABk
In upper right comer region as FABkIn the 2nd region be denoted asBy FABkIn lower left corner region as FABkIn
3rd region is denoted asBy FABkIn lower right field as FABkIn the 4th region be denoted asWherein, I
Indicate that the horizontal direction resolution ratio of F, J indicate the vertical direction resolution ratio of F, symbolFor downward rounding operation symbol, k is
Positive integer,
1. _ 2, to each of FA each region in the block step carries out discrete cosine transform, obtain each of FA
The dct coefficient matrix in son each region in the block, willDct coefficient matrix be denoted asIt willDct coefficient matrix be denoted asIt willDct coefficient matrix note
ForIt willDct coefficient matrix be denoted asWherein, WithDimension be 4 × 4.
Step 1. _ 3, create feature watermark W, detailed process is：Calculate the discrete remaining of each sub-block in FA
String converts DC coefficient mean value, by FABkDiscrete cosine transform DC coefficient mean value be denoted as FADk,Then the discrete cosine transform DC coefficient according to all sub-blocks in FA is equal
Value, establishment obtain feature watermark W, the kth position bit of W are denoted as Wk,Wherein, the length of W is It indicatesIn the 1st column element of the 1st row,ForDiscrete cosine transform
DC coefficient,It indicatesIn the 1st column element of the 1st row,ForDiscrete cosine transform it is straight
Flow coefficient, δ1For the first threshold of setting, δ is taken in the present embodiment1=1000.
Step 1. _ 4, self-embedding feature watermark W, detailed process is：Self-embedding feature water in each sub-block in FA
A bit in W is printed, for FABk, take out WkIf Wk=0 andThen enableThen it enablesExchangeWithValue, realize WkSelf-embedding, then in order to improve the robustness of embedded digital watermarking, whenIt is right when establishmentWithIt is for further adjustments, i.e.,
It enablesAnd it enables
If Wk=0 andIt is then rightWithIt does not deal with, realizes
WkSelf-embedding, then in order to improve the robustness of embedded digital watermarking, when
It is right when establishmentWithIt is for further adjustments, evenAnd it enables
If Wk=1 andThen enable
Then it enablesExchangeWithValue, realize Wk's
Self-embedding, then in order to improve the robustness of embedded digital watermarking, whenAt
Immediately, rightWithIt is for further adjustments, evenAnd it enables
If Wk=1 andIt is then rightWithIt does not deal with, realizes
WkSelf-embedding, then in order to improve the robustness of embedded digital watermarking, when
It is right when establishmentWithIt is for further adjustments, evenAnd it enables
Wherein,It indicatesIn the 2nd column element of the 2nd row,It indicatesIn the 2nd
The 2nd column element of row, G1 and G2 are the intermediate variable of introducing, and abs () is the function that takes absolute value, δ2For the second threshold of setting,
δ is taken in the present embodiment2=20, In "=" be assignment.
Step 1. _ 5, step 1. _ 4 on the basis of, to each region carry out inverse discrete cosine transform；Then recombination obtains
Each sub-block for being embedded with digital watermarking, and then recombinate and obtain the ll channel embedded with digital watermarking；Then to being embedded with digital watermarking
Ll channel and the three width details subgraphs of F carry out level-one inverse discrete wavelet transform, obtain the watermarking images embedded with digital watermarking.
The digital watermarking extraction and detection part the specific steps are：
Step 2. _ 1, the watermarking images embedded with digital watermarking be 8bit gray level images, be denoted as TF, the size of TF is I
×J；Then level-one discrete wavelet transformation is carried out to TF, the three width details subgraphs and a width ll channel of TF is obtained, by forcing for TF
Nearly subgraph is denoted as TFA, and the size of TFA isThen TFA is divided by raster scan orderIt is a mutual
The sub-block that nonoverlapping size is 8 × 8, TFAB is denoted as by k-th of sub-block in TFAk；Again by each sub-block in TFA
It is divided into the region that the size of 4 non-overlapping copies is 4 × 4, by TFABkIn upper left corner area as TFABkIn the 1st
A region is denoted asBy TFABkIn upper right comer region as TFABkIn the 2nd region be denoted asBy TFABk
In lower left corner region as TFABkIn the 3rd region be denoted asBy TFABkIn lower right field as TFABk
In the 4th region be denoted asWherein, symbolFor downward rounding operation symbol, k is positive integer,
2. _ 2, to each of TFA each region in the block step carries out discrete cosine transform, obtain every in TFA
The dct coefficient matrix in height each region in the block, willDct coefficient matrix be denoted asIt willDct coefficient matrix be denoted asIt willDiscrete cosine transform coefficient square
Battle array is denoted asIt willDct coefficient matrix be denoted asWherein, WithDimension be 4 × 4.
Step 2. _ 3, Blind extracting feature watermark TW, detailed process is：The discrete cosine for calculating each sub-block in TFA becomes
DC coefficient mean value is changed, by TFABkDiscrete cosine transform DC coefficient mean value be denoted as TFADk,Then according to the discrete cosine transform direct current system of all sub-blocks in TFA
Number mean value, Blind extracting obtain feature watermark TW, the kth position bit of TW are denoted as TWk,Its
In, the length of TW is It indicatesIn the 1st column element of the 1st row,ForDiscrete cosine transform DC coefficient,It indicatesIn the 1st column element of the 1st row,
ForDiscrete cosine transform DC coefficient, δ1For the first threshold of setting, δ is taken in the present embodiment1=1000.
Step 2. _ 4, Blind extracting authenticating water-mark TW*, detailed process is：One is blindly extracted in each sub-block in TFA
Bit is to obtain authenticating water-mark TW*, for TFABk, extract and obtain TW*In kth position bit TW* k,Wherein, TW*Length be Table
ShowIn the 2nd column element of the 2nd row,It indicatesIn the 2nd column element of the 2nd row.
Step 2. _ 5, according to TW and TW*Copyright protection is carried out, detailed process is：Calculate TW and TW*Between normalization phase
Relationship number is denoted as ρ (TW, TW*),Wherein,Indicate the average value of all elements in TW,Indicate TW*In all elements average value；Utilize ρ (TW, TW*) come
Carry out copyright protection.
According to TW and TW*Tampering location is carried out, detailed process is：According to the every bit and TW in TW*In corresponding ratio
Spy, determines whether each sub-block in TFA is tampered with, for TFABkIf TWkWith TW* kIt is equal, it is determined that TFABkIt is not usurped
It corrects one's mistakes；If TWkWith TW* kIt is unequal, it is determined that TFABkIt is tampered with, realizes total blindness's tampering location.
Method is applied to the feasibility in terms of digital image copyright protection and tampering location in order to better illustrate the present invention,
Experiment simulation is carried out using the original Lena gray level images that resolution ratio shown in FIG. 1 is 512 × 512 as test image, experiment is imitative
Really carried out on Matlab7.5 platforms.Wherein, first threshold δ1Value is 1000, second threshold δ2Value is 20.Fig. 2 gives
The watermark embedded with watermark obtained after being handled original Lena gray level images shown in FIG. 1 using the method for the present invention is gone out
Lena gray level images, from fig. 2 it can be seen that the masters of watermark Lena gray level images, objective quality embedded with watermark are very good,
Y-PSNR has reached 38.23dB, is difficult visually to feel variation, meets the requirement of watermark imperceptible.Using this
Normalizated correlation coefficient between the feature watermark and authenticating water-mark of inventive method extraction is 1, illustrates the watermark embedded with watermark
Lena gray level images can extract embedded feature watermark completely losslessly when not destroyed by any processing or attack, because
This can be used for the copyright protection of digital picture works.
Here, the picture quality after digital picture insertion watermark, and picture quality after image procossing are objective
Evaluation is all made of Y-PSNR (PSNR),Wherein, I=
Coordinate position is the pixel of (i, j) in 512, J=512,1≤i≤I, 1≤j≤J, f (i, j) expression original-gray image F
Pixel value, tf (i, j) is indicated after embedded watermark or watermarking images coordinate bit in gray level image TF by image procossing and after distorting
It is set to the pixel value of the pixel of (i, j), fmaxIndicate that the max pixel value in original-gray image F, I × J indicate original gradation
The resolution ratio of image F.
Below by way of to embedded with watermark watermark Lena gray level images carry out it is various processing and distort operation, to verify this hair
Bright method is for the feasibility in terms of digital image copyright protection and tampering location.
1) histogram equalization
Histogram equalization processing is carried out to watermark Lena gray level images shown in Fig. 2, is obtained through at histogram equalization
Watermark Lena gray level images after reason, as shown in Figure 3.After histogram equalization processing, in watermark Lena gray level images
Significant change has occurred in the pixel Distribution value of pixel, and Y-PSNR (PSNR) is only 19.04dB.But use the method for the present invention
Normalizated correlation coefficient between the feature watermark and authenticating water-mark that extract has reached 0.89, has absolutely proved embedded feature
Watermark can be extracted well, illustrate the robustness of the method for the present invention.
2) JPEG lossy compressions
JPEG lossy compression processing is carried out to watermark Lena gray level images shown in Fig. 2, compression quality selecting predictors are
5%, the watermark Lena gray level images after JPEG lossy compressions are obtained, as shown in Figure 4.Figure 4, it is seen that having through JPEG
It damages compressed watermark Lena gray level images and shows certain blocking artifact, visual quality is degenerated, Y-PSNR
(PSNR) 33.62dB is dropped to.But the normalization phase between the feature watermark and authenticating water-mark extracted using the method for the present invention
Relationship number has reached 0.83, has absolutely proved that embedded feature watermark can be extracted well, has illustrated the method for the present invention
Robustness.
3) it is superimposed Gauss (Gauss) noise
Noise overlap-add procedure is carried out to watermark Lena gray level images shown in Fig. 2, it is 0 and variance that the noise of superposition, which is mean value,
For 0.001 Gaussian reflectivity mirrors, the watermark Lena gray level images after superimposed Gaussian noise are obtained, as shown in Figure 5.From Fig. 5
In as can be seen that more serious degeneration, peak value occur for the visual quality of the watermark Lena gray level images after superimposed Gaussian noise
Signal-to-noise ratio (PSNR) is only 29.41dB.But the normalizing between the feature watermark and authenticating water-mark extracted using the method for the present invention
Changing related coefficient has 0.96, illustrates that embedded feature watermark almost do not influenced by.
4) medium filtering
Median filter process, the window size selection of median filter are carried out to watermark Lena gray level images shown in Fig. 2
For [5 × 5], the watermark Lena gray level images after medium filtering are obtained, as shown in Figure 6.From fig. 6 it can be seen that through intermediate value
The detailed information of filtered watermark Lena gray level images occurs to obscure, and Y-PSNR (PSNR) drops to 32.89dB.But it adopts
Normalizated correlation coefficient between the feature watermark and authenticating water-mark that are extracted with the method for the present invention still has 0.86, illustrates insertion
Feature watermark remains to preferably be extracted the copyright protection for digital picture works.
The multifunctionality of method in order to further illustrate the present invention dislikes watermark Lena gray level images shown in Fig. 2
Tampering detection of anticipating is tested, and the sub-block of distorting detected is shown with density bullet.The malice of image is distorted main including replacing, cutting
It cuts and the operations such as pixel value increase.
Region replacement operation 1：By (the 109 of watermark Lena gray level images shown in Fig. 2:236,149:276) region is replaced
The baboon gray level images for being 128 × 128 at size shown in Fig. 7, the watermark Lena gray-scale maps obtained after replacement operation
As shown in Figure 8.The tampering location testing result obtained using the method for the present invention is usurped as shown in figure 9, can orient well
The image-region changed.
Region replacement operation 2：By (the 129 of watermark Lena gray level images shown in Fig. 2:256,129:256) in region
Pixel is substituted for another region (229 of images themselves:356,229:356) pixel in, the water obtained after replacement operation
It is as shown in Figure 10 to print Lena gray level images.The tampering location testing result obtained using the method for the present invention is as shown in figure 11, can be with
The image-region distorted is oriented well.
Region replacement operation 3：By (the 329 of watermark Lena gray level images shown in Fig. 2:456,229:356) region is replaced
At (the 229 of original Lena gray level images shown in FIG. 1:356,379:506) region, the watermark Lena ashes obtained after replacement operation
It is as shown in figure 12 to spend image.The tampering location testing result obtained using the method for the present invention is as shown in figure 13, can accurately determine
Position goes out the image-region distorted.
Region cutting operation：Watermark Lena gray level images shown in Fig. 2 are located at (214:267,189:227) in region
Image block is cut away, i.e., the pixel value of each pixel in the region is set to 0, obtains the watermark after the cutting operation of region
Lena gray level images are as shown in figure 14.The tampering location testing result obtained using the method for the present invention is as shown in figure 15, Ke Yizhun
Really detect and orient the image-region distorted.
Pixel value increases operation：(329 are located to watermark Lena gray level images shown in Fig. 2:456,129:256) in region
Image block into row pixel value increase operate, such as by the pixel value of each pixel in the region add 50, obtain pixel
The watermark Lena gray level images that value increases after operation are as shown in figure 16.The tampering location testing result obtained using the method for the present invention
As shown in figure 17, the image-region distorted can be oriented well.
Claims (3)
1. a kind of total blindness's digital watermark method with copyright protection and tampering location function, it is characterised in that main processes
For：
At digital watermark embedding end, level-one discrete wavelet transformation, the approximator that will be obtained are carried out to original 8bit gray level images
Figure be divided into multiple non-overlapping copies size be 8 × 8 sub-block, by each sub-block in ll channel be divided into 4 mutually
The region that nonoverlapping size is 4 × 4；Then each of ll channel each region in the block is carried out discrete remaining
String converts；The discrete cosine transform direct current of upper left corner area and upper right comer region then according to all sub-blocks in ll channel
Coefficient, establishment obtain feature watermark；Lower left corner region and lower right field further according to each sub-block in ll channel it is same
Every bit in the feature watermark that the discrete cosine transform ac coefficient of one position and establishment obtain, it is every in ll channel
A bit in a sub-block in self-embedding feature watermark；Finally by inverse discrete cosine transform and level-one inverse discrete wavelet transform
Obtain the watermarking images embedded with digital watermarking；
In digital watermarking extraction and test side, level-one discrete wavelet transformation is carried out to the watermarking images embedded with digital watermarking, will
To ll channel be divided into the sub-block that the size of multiple non-overlapping copies is 8 × 8, by each sub-block in ll channel point
It is cut into the region that the size of 4 non-overlapping copies is 4 × 4；Then to each of ll channel each region in the block
Carry out discrete cosine transform；Upper left corner area and upper right comer region then according to all sub-blocks in ll channel it is discrete remaining
String converts DC coefficient, and Blind extracting obtains feature watermark；Lower left corner region further according to each sub-block in ll channel and the right side
The discrete cosine transform ac coefficient of the same position of lower angular zone, Blind extracting obtain authenticating water-mark；It is carried finally by calculating is blind
Normalizated correlation coefficient between the authenticating water-mark that the feature watermark and Blind extracting obtained obtains carries out copyright protection, passes through
Authenticating water-mark that the feature watermark and Blind extracting that successive appraximation Blind extracting obtains obtain realizes total blindness's tampering location.
2. a kind of total blindness's digital watermark method with copyright protection and tampering location function according to claim 1,
It is characterized in that total blindness's digital watermark method includes two digital watermark embedding, digital watermarking extraction and detection parts；
The digital watermark embedding part the specific steps are：
Step 1. _ 1, choose an original 8bit gray level image, and be denoted as F, the size of F is I × J；Then F is carried out
Level-one discrete wavelet transformation obtains the three width details subgraphs and a width ll channel of F, the ll channel of F is denoted as FA, the ruler of FA
Very little size isThen FA is divided intoThe sub-block that the size of a non-overlapping copies is 8 × 8, will be in FA
K-th of sub-block be denoted as FABk；Each sub-block in FA is divided into the area that the size of 4 non-overlapping copies is 4 × 4 again
Domain, by FABkIn upper left corner area as FABkIn the 1st region be denoted asBy FABkIn upper right comer region conduct
FABkIn the 2nd region be denoted asBy FABkIn lower left corner region as FABkIn the 3rd region be denoted as
By FABkIn lower right field as FABkIn the 4th region be denoted asWherein, I indicates that the horizontal direction of F is differentiated
Rate, J indicate the vertical direction resolution ratio of F, symbolFor downward rounding operation symbol, k is positive integer,
1. _ 2, to each of FA each region in the block step carries out discrete cosine transform, obtain each sub-block in FA
In each region dct coefficient matrix, willDct coefficient matrix be denoted asIt willDct coefficient matrix be denoted asIt willDct coefficient matrix be denoted as
It willDct coefficient matrix be denoted asWherein, WithDimension
It is 4 × 4；
Step 1. _ 3, create feature watermark W, detailed process is：Calculate the discrete cosine transform direct current system of each sub-block in FA
Number mean value, by FABkDiscrete cosine transform DC coefficient mean value be denoted as FADk,
Then according to the discrete cosine transform DC coefficient mean value of all sub-blocks in FA, establishment obtains feature watermark W, by the kth position of W
Bit is denoted as Wk,Wherein, the length of W is It indicatesIn
The 1st column element of the 1st row,ForDiscrete cosine transform DC coefficient,It indicatesIn
The 1st column element of 1st row,ForDiscrete cosine transform DC coefficient, δ1For the first threshold of setting；
Step 1. _ 4, self-embedding feature watermark W, detailed process is：A ratio in each sub-block in FA in self-embedding feature watermark W
Spy, for FABk, take out WkIf Wk=0 andThen enable
Then it enablesRealize WkSelf-embedding, then when
It is right when establishmentWithIt is for further adjustments, evenAnd
It enables
If Wk=0 andIt is then rightWithIt does not deal with, realizes Wk's
Self-embedding, then whenIt is right when establishmentWithMake
Further adjustment, evenAnd it enables
If Wk=1 andThen enableThen it enablesRealize WkSelf-embedding, then when
It is right when establishmentWithIt is for further adjustments, evenAnd it enables
If Wk=1 andIt is then rightWithIt does not deal with, realizes Wk's
Self-embedding, then whenIt is right when establishmentWithMake
Further adjustment, evenAnd it enables
Wherein,It indicatesIn the 2nd column element of the 2nd row,It indicatesIn the 2nd row the 2nd
Column element, G1 and G2 are the intermediate variable of introducing, and abs () is the function that takes absolute value, δ2For the second threshold of setting, In "=" be assignment；
Step 1. _ 5, step 1. _ 4 on the basis of, to each region carry out inverse discrete cosine transform；Then recombination obtains each
Sub-block embedded with digital watermarking, and then recombinate and obtain the ll channel embedded with digital watermarking；Then to being forced embedded with digital watermarking
The three width details subgraphs of nearly subgraph and F carry out level-one inverse discrete wavelet transform, obtain the watermarking images embedded with digital watermarking；
The digital watermarking extraction and detection part the specific steps are：
Step 2. _ 1, the watermarking images embedded with digital watermarking be 8bit gray level images, be denoted as TF, the size of TF is I × J；
Then level-one discrete wavelet transformation is carried out to TF, the three width details subgraphs and a width ll channel of TF is obtained, by the approximator of TF
Seal is TFA, and the size of TFA isThen TFA is divided intoThe size of a non-overlapping copies is
K-th of sub-block in TFA is denoted as TFAB by 8 × 8 sub-blockk；Each sub-block in TFA is divided into 4 non-overlapping copies again
The region that size is 4 × 4, by TFABkIn upper left corner area as TFABkIn the 1st region be denoted asIt will
TFABkIn upper right comer region as TFABkIn the 2nd region be denoted asBy TFABkIn lower left corner region conduct
TFABkIn the 3rd region be denoted asBy TFABkIn lower right field as TFABkIn the 4th region be denoted asWherein, symbolFor downward rounding operation symbol, k is positive integer,
2. _ 2, to each of TFA each region in the block step carries out discrete cosine transform, obtain every height in TFA
The dct coefficient matrix in each region in the block, willDct coefficient matrix be denoted asIt willDct coefficient matrix be denoted asIt willDiscrete cosine transform coefficient square
Battle array is denoted asIt willDct coefficient matrix be denoted asWherein, WithDimension be 4 × 4；
Step 2. _ 3, Blind extracting feature watermark TW, detailed process is：Calculate the discrete remaining of each sub-block in TFA
String converts DC coefficient mean value, by TFABkDiscrete cosine transform DC coefficient mean value be denoted as TFADk,Then according to the discrete cosine transform direct current system of all sub-blocks in TFA
Number mean value, Blind extracting obtain feature watermark TW, the kth position bit of TW are denoted as TWk,Its
In, the length of TW is It indicatesIn the 1st column element of the 1st row,ForDiscrete cosine transform DC coefficient,It indicatesIn the 1st column element of the 1st row,
ForDiscrete cosine transform DC coefficient, δ1For the first threshold of setting；
Step 2. _ 4, Blind extracting authenticating water-mark TW*, detailed process is：A bit is blindly extracted in each sub-block in TFA
To obtain authenticating water-mark TW*, for TFABk, extract and obtain TW*In kth position bit TW* k,Wherein, TW*Length be Table
ShowIn the 2nd column element of the 2nd row,It indicatesIn the 2nd column element of the 2nd row；
Step 2. _ 5, according to TW and TW*Copyright protection is carried out, detailed process is：Calculate TW and TW*Between normalization phase relation
Number, is denoted as ρ (TW, TW*), utilize ρ (TW, TW*) carry out copyright protection；
According to TW and TW*Tampering location is carried out, detailed process is：According to the every bit and TW in TW*In corresponding bit, really
Whether each sub-block determined in TFA is tampered with, for TFABkIf TWkWith TW* kIt is equal, it is determined that TFABkIt was not tampered with；
If TWkWith TW* kIt is unequal, it is determined that TFABkIt is tampered with, realizes total blindness's tampering location.
3. a kind of total blindness's digital watermark method with copyright protection and tampering location function according to claim 2,
In being characterized in that the step 2. _ 5,Its
In,Indicate the average value of all elements in TW,Indicate TW*In all elements average value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810299604.9A CN108648130B (en) | 2018-04-04 | 2018-04-04 | Totally-blind digital watermarking method with copyright protection and tampering positioning functions |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810299604.9A CN108648130B (en) | 2018-04-04 | 2018-04-04 | Totally-blind digital watermarking method with copyright protection and tampering positioning functions |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108648130A true CN108648130A (en) | 2018-10-12 |
| CN108648130B CN108648130B (en) | 2022-05-27 |
Family
ID=63745452
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810299604.9A Active CN108648130B (en) | 2018-04-04 | 2018-04-04 | Totally-blind digital watermarking method with copyright protection and tampering positioning functions |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108648130B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109977878A (en) * | 2019-03-28 | 2019-07-05 | 华南理工大学 | Vehicle checking method based on weight A weighting nchor |
| CN110555792A (en) * | 2019-08-16 | 2019-12-10 | 广东外语外贸大学南国商学院 | Image tampering blind detection method based on normalized histogram comprehensive feature vector |
| CN110580405A (en) * | 2019-08-16 | 2019-12-17 | 湖北工业大学 | Picture copyright protection system and method based on social network |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102184520A (en) * | 2011-05-26 | 2011-09-14 | 浙江工商大学 | Complete blind detection robust quantization watermarking method |
| CN102208097A (en) * | 2011-05-26 | 2011-10-05 | 浙江工商大学 | Network image copyright real-time distinguishing method |
| CN102223561A (en) * | 2011-07-01 | 2011-10-19 | 宁波大学 | Blind watermark embedding and extracting method of stereoscopic video image |
| EP2428930A1 (en) * | 2009-09-03 | 2012-03-14 | ZTE Corporation | Method and system for embedding and extracting image digital watermark |
| CN102682418A (en) * | 2012-05-11 | 2012-09-19 | 宁波大学 | Method for embedding and extracting multiple zero watermarks of digital image |
-
2018
- 2018-04-04 CN CN201810299604.9A patent/CN108648130B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP2428930A1 (en) * | 2009-09-03 | 2012-03-14 | ZTE Corporation | Method and system for embedding and extracting image digital watermark |
| CN102184520A (en) * | 2011-05-26 | 2011-09-14 | 浙江工商大学 | Complete blind detection robust quantization watermarking method |
| CN102208097A (en) * | 2011-05-26 | 2011-10-05 | 浙江工商大学 | Network image copyright real-time distinguishing method |
| CN102223561A (en) * | 2011-07-01 | 2011-10-19 | 宁波大学 | Blind watermark embedding and extracting method of stereoscopic video image |
| CN102682418A (en) * | 2012-05-11 | 2012-09-19 | 宁波大学 | Method for embedding and extracting multiple zero watermarks of digital image |
Non-Patent Citations (3)
| Title |
|---|
| 薛栋，周亚训，金炜: "版权保护和内容认证的全盲双功能数字水印算法", 《研究与开发》 * |
| 薛栋: "基于自嵌入技术的图像内容认证和恢复的研究", 《中国优秀博硕士学位论文全文数据库(硕士)信息科技辑（月刊）》 * |
| 赵伟;肖世校;陈仁安: "基于ICA的混合域盲水印算法", 《实验室研究与探索》 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109977878A (en) * | 2019-03-28 | 2019-07-05 | 华南理工大学 | Vehicle checking method based on weight A weighting nchor |
| CN110555792A (en) * | 2019-08-16 | 2019-12-10 | 广东外语外贸大学南国商学院 | Image tampering blind detection method based on normalized histogram comprehensive feature vector |
| CN110580405A (en) * | 2019-08-16 | 2019-12-17 | 湖北工业大学 | Picture copyright protection system and method based on social network |
| CN110555792B (en) * | 2019-08-16 | 2022-05-17 | 广东外语外贸大学南国商学院 | Image tampering blind detection method based on normalized histogram comprehensive feature vector |
| CN110580405B (en) * | 2019-08-16 | 2022-07-05 | 湖北工业大学 | Picture copyright protection system and method based on social network |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108648130B (en) | 2022-05-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US7995790B2 (en) | Digital watermark detection using predetermined color projections | |
| Chou et al. | A perceptually tuned watermarking scheme for color images | |
| US6631198B1 (en) | Perceptual modeling of media signals based on local contrast and directional edges | |
| CN112241929B (en) | Watermark embedding method, watermark detecting method, watermark embedding device, watermark detecting device and electronic equipment | |
| CN101246588B (en) | Self-adapting watermarking algorithm of colorful image hypercomplex number spacing | |
| CN110599388B (en) | Blind robust digital watermark embedding and detecting method based on positioning point assistance | |
| CN108416727A (en) | A kind of coloured image total blindness's robust digital watermark method based on self-embedding technology | |
| CN108648130A (en) | A kind of total blindness's digital watermark method with copyright protection and tampering location function | |
| CN102880999A (en) | Color image digital watermark method capable of resisting stamping, printing and scanning processes | |
| CN102024249A (en) | Digital image watermarking method based on visual perception characteristics | |
| Khalifa et al. | A robust non-blind algorithm for watermarking color images using multi-resolution wavelet decomposition | |
| CN110189243A (en) | A kind of color image robust watermarking method based on tensor singular value decomposition | |
| Riad et al. | Robust fourier watermarking for id images on smart card plastic supports | |
| Nguyen et al. | Perceptual watermarking using a new Just-Noticeable-Difference model | |
| Fung et al. | A novel DWT-SVD video watermarking scheme using side view | |
| CN110246076A (en) | The high dynamic range images water mark method decomposed based on Tucker | |
| Fung et al. | A new approach of DWT-SVD video watermarking | |
| CN1329874C (en) | Universal digital image invisible information detecting method | |
| Rawat et al. | DCT and SVD based new watermarking scheme | |
| Liu et al. | Optimization-based image watermarking algorithm using a maximum-likelihood decoding scheme in the complex wavelet domain | |
| Abdel-Aziz et al. | On perceptual quality of watermarked images–an experimental approach | |
| Masry et al. | Digital watermarking using local contrast-based texture masking | |
| Yaghmaee et al. | Computing watermark capacity in images according to their quad tree | |
| Al-Otum et al. | Color image watermarking based on self-embedded color permissibility with preserved high image quality and enhanced robustness | |
| Hosam et al. | A hybrid ROI-embedding based watermarking technique using DWT and DCT transforms |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |