CN1885339A - Electronic stamp certification method based on image features - Google Patents

Abstract

The invention relates to an electric stamp checking method, based on image character, wherein said invention combines the image character of stamp and the semi-fragility watermark; the invention first pretreats the original stank image, then extracts the character of stamp as the watermark to be embedded into stamp image; when the stamp needs identification, extracting watermark, to judge if the error between the watermark and the character of checked stamp is in acceptable range, to confirm if the stamp is reliable; if it is not, confirming the position that changed maliciously. The invention can correctly judge the treatment on stamp is malicious or normal, and accurately position the changed position, belonging to the multimedia signal processing technique.

Description

A kind of electronic stamp certification method based on characteristics of image

Technical field

The invention belongs to the multimedia signal processing technique field, be specifically related to a kind of in conjunction with seal image feature and semi-fragile watermarking, can correctly judge to the E-seal treatment of picture it is that malice is distorted or normal Flame Image Process, and can be to the pinpoint electronic stamp certification method in the position of distorting.

Background technology

E-seal is to develop and the new technology of proposition in seasonable generation.It not only can remedy the deficiency and the defective of traditional seal, and along with the develop rapidly of e-commerce and e-government, the use of E-seal will be more and more widely.Yet seal may be through distorting of having a mind in network transmission process, so its authenticity must obtain confirming in use, otherwise can bring very serious consequence.How in network environment, authenticity, the integrality of seal content are implemented effective protection has become a stern reality problem.

Traditional authentication techniques based on cryptographic digital signature are applied to seal image and come with some shortcomings.For example, too responsive to normal signal Processing (as data compression, signal filtering etc.) and interchannel noise (as wireless channel etc.), can't the positioning tampering position etc.And E-seal can be compressed processing or interference of noise usually, distorts in the sensitivity for malice when authenticating, and tackles above-mentioned normal picture processing and has certain tolerance.Can accomplish to distort sensitivity and the accurate positioning tampering position of energy for malice based on the image authentication technology of digital watermarking, simultaneously, be robust to normal signal Processing or noise, has stronger practicality.Adopt and watermark that image is irrelevant at present morely, be subjected to the attack of pseudo-authentication easily, and bearing accuracy and robustness aspect await raising based on the authentication method of watermark.

Summary of the invention

For overcoming the deficiencies in the prior art, the present invention proposes a kind of electronic stamp certification method based on characteristics of image.The present invention utilizes the feature of seal image itself to be embedded in the seal image as watermark, when what seal image was carried out is that normal picture is when handling, think during authentication that seal image is believable, and when seal image is subjected to malice and distorts, can judge that not only seal image is incredible, but also position that can positioning tampering.

The present invention realizes that the technical scheme of above-mentioned purpose is as follows: method of the present invention is divided into watermark embedding and two processes of authentication seal;

The watermark embed process step is as follows: 1) original seal is carried out pre-service; 2) characteristics of image of the original seal of extraction; 3) pre-service watermark; 4) structuring embed watermark;

The authentication seal process steps is as follows: 1) extract seal image feature to be certified; 2) characteristics of image of the original seal of reconstruct; 3) seal to be certified is authenticated.

Described watermark embed process step 1) is carried out pre-service to original seal and is adopted following formula:

$X^{\&prime;}=\left\{\begin{array}{cc}X& 0\&le;X\&le;T\\ T& T<X\&le;255\end{array}\right.$

Wherein, X represents original seal image pixel value, and X ' is the pixel value of seal image after the pre-service, and T is pretreated threshold value, gets 245.

Described watermark embed process step 2) the characteristics of image method of the original seal of extraction is as follows: utilize the edge of edge extracting algorithm extraction seal figure and carry out binaryzation, on binary image, calculate the amplitude of 49 Zernike squares, be designated as ZMM, and it is normalized to decimal between the 0-1; Each ZMM is quantified as 16, get the highest 4 as watermark.

Described watermark embed process step 3) pretreated water impression method is as follows: the watermark bit sequence W length of establishing embedding is N, generate 0, the 1 random series R that length is N by secret key k, to watermark sequence W and immediately sequence R carry out xor operation generation watermark information W to be embedded ^*

Described watermark embed process step 4) structuring embed watermark method is as follows: at first seal image is carried out 3 grades of two-dimensional discrete wavelet conversions, obtain 10 subband { LH _i, HL _i, HH _i, LL ₃, i=1-3}, wherein LL ₃Subband is that the low pass of original seal image is approximate, has concentrated most of energy of seal image signal, LL behind 3 grades of wavelet transformations ₃8 * 8 piece is relevant in wavelet coefficient values of subband and the real image, LL ₃The piece of subband 2 * 2 is relevant with 16 * 16 of real images; Successively high 4 forms with 2 * 2 piece of each ZMM are embedded into LL then ₃In the subband, 16 * 16 piece is distorted the value that may change a square in real image, can determine the position that malice is distorted according to the position of the square that has changed in when authentication; Carry out wavelet inverse transformation at last, obtain the seal image of embed watermark.

Described authentication seal process steps 1) it is as follows to extract seal image characterization method to be certified: the edge and the binaryzation of getting seal image to be certified, on binaryzation edge seal, extract the amplitude of 49 Zernike squares, be called ZMM_A, 49 ZMM_A have represented the feature of image to be certified.

Described authentication seal process steps 2) the characteristics of image method of the original seal of reconstruct is as follows: at first treat authentication image and carry out 3 grades of two-dimensional wavelet transformations, obtain 10 subband { LH _i, HL _i, HH _i, LL ₃, i=1-3}; In each piece of 2 * 2 of LL3 subband, extract watermark then, obtain watermark information W ', by identical secret key k in the watermark pretreatment when embedding, generate random series R, to watermark sequence W ' and immediately sequence R carry out xor operation and generate watermark information W "; " amplitude of 49 Zernike squares of the original seal image of reconstruct is called ZMM_O by W.

Described authentication seal process steps 3) it is as follows seal to be certified to be carried out authentication method: the quadratic sum of calculating the difference of 49 corresponding ZMM_A and ZMM_O, and judge whether to have surpassed preset threshold 0.125, if it is littler than threshold value, think that then seal is believable, otherwise think that this seal suffered malicious attack, further with the difference of each ZMM_A and ZMM_O square with selected threshold value 0.003 relatively, if greater than threshold value, think that then the embedded location place of this ZMM correspondence has suffered that malice distorts, 16 * 16 of corresponding spatial domain.

The embedding grammar of each watermark is as follows in the described watermark embed process step 4):

Define two variable S_LSB6 (a, b) and G_LSB6 (a), wherein S_LSB6 (a, b) expression is with low six low six of replacing a of b, and G_LSB6 (a) expression is taken out low six of a, when the embed watermark bit be ' 1 ', the use formula:

$F_i^{\&prime;}(u,v)=\left\{\begin{array}{cc}S\_\mathrm{LSB}6(F_i(u,v)-010000b,110000b)& G\_\mathrm{LSB}6\left(F_i(u,v)\right)\&le;010000b\\ S\_\mathrm{LSB}6(F_i(u,v),110000)& \mathrm{otherwise}\end{array}\right.$

When the embed watermark bit is ' 0 ', use formula

$F_i^{\&prime;}(u,v)=\left\{\begin{array}{cc}S\_\mathrm{LSB}6(F_i(u,v)+100000b,010000b)& G\_\mathrm{LSB}6\left(F_i(u,v)\right)\&le;110000b\\ S\_\mathrm{LSB}6(F_i(u,v),010000b)& \mathrm{otherwise}\end{array}\right.$

Wherein, F _i(u v) is small echo LL ₃Sub-band coefficients, F _i' (u v) is that watermark embeds the back coefficient.

Described authentication seal process steps 2) extracting method of watermark bit is to obtain LL in ₃The 6th least significant bit (LSB) of the binary representation of the integral part of sub-band coefficients.

The present invention relatively has following advantage with existing electron seal identification technology based on watermark:

1) feature with seal image itself embeds seal image as watermark, and utilizes the difference between the feature to authenticate, and makes authentication result more reliable.

2) seal image is carried out pre-service, the reduction pixel is overflowed the influence to the watermark robustness, improves the accuracy of authentication.

3) the algorithm invisibility of the present invention's proposition is good, and is better to the robustness that normal image is handled, and malice is distorted the comparison sensitivity, and bearing accuracy is higher.The Y-PSNR of seal image behind the inventive method embed watermark, note is made PSNR, and its value is for more than the 42dB, the robustness of antagonism JPEG compression reaches compressibility factor 40%, to antimierophonic intensity is S=4, and S=4 represents the random noise between the original seal image stack [20,20].When image generation content tampering, authentication will think that this seal image is incredible, and the bearing accuracy of distorting is 16 * 16 piece.

Description of drawings

Fig. 1 carries out 3 grades of 2-d wavelets to seal image in the seal verifying method that proposes of the present invention to decompose the subband synoptic diagram that obtains;

Fig. 2 is that the structuring of watermark in the seal verifying method that proposes of the present invention embeds synoptic diagram;

To be the authentication method that proposes of the present invention carry out the synoptic diagram that pre-service and watermark embed to seal 1 to Fig. 3;

To be the authentication method that proposes of the present invention carry out the synoptic diagram that pre-service and watermark embed to seal 2 to Fig. 4;

To be the authentication method that proposes of the present invention carry out the synoptic diagram that pre-service and watermark embed to seal 3 to Fig. 5;

To be the authentication method that proposes of the present invention carry out the synoptic diagram that pre-service and watermark embed to seal 4 to Fig. 6;

Fig. 7 is the authentication result of authentication method when seal 1 figure is distorted by malice that the present invention proposes;

Fig. 8 is the authentication result of authentication method when seal 2 figure are distorted by malice that the present invention proposes;

Fig. 9 is the authentication result of authentication method when seal 3 figure are distorted by malice that the present invention proposes;

Figure 10 is the authentication result of authentication method when seal 4 figure are distorted by malice that the present invention proposes.

Embodiment

Below in conjunction with accompanying drawing the inventive method is described further.

The inventive method is divided into watermark embedding and two processes of authentication seal, and the watermark embed process step is as follows:

1. original seal is carried out pre-service

Overflowing of pixel is one of factor that influences the watermark robustness.In general, the shared ratio of contiguous pixels is big more in the image, and the probability that pixel is overflowed behind the embed watermark is high more, and then the influence that is subjected to of the robustness of watermark is big more.Can see that from histogrammic statistics near the ratio of the pixel of the gray scale of seal image 255 is very big, overflows easily in watermark embed process and image processing process.The present invention is to the pre-service of seal image formula as follows.

$X^{\&prime;}=\left\{\begin{array}{cc}X& 0\&le;X\&le;T\\ T& T<X\&le;255\end{array}\right.---\left(1\right)$

Wherein, X represents original seal image pixel value, and X ' is the pixel value of seal image after the pre-service, and T is pretreated threshold value, gets 245.Because near the pixel 255 is background mostly, pre-service can not influence the content of seal.

2. extract the characteristics of image of original seal

Utilize the edge extracting algorithm to extract the edge of seal figure and carry out binaryzation, on binary image, calculate the amplitude of 49 Zernike squares, be designated as ZMM, and it is normalized to decimal between the 0-1.The Zernike square is proposed in 1934 by Zernike the earliest, and its computing method are as follows:

For digital picture f (x, y), n rank multiplicity is that the Zernike square of m is defined as:

$A_{\mathrm{nm}}=\frac{n+1}{\mathrm{\&pi;}}\underset{x}{\mathrm{\&Sigma;}}\underset{y}{\mathrm{\&Sigma;}}f(x,y)V_{\mathrm{nm}}^{*}(\mathrm{\&rho;},\mathrm{\&theta;}),x^2+y^2\&le;1---\left(1\right)$

Wherein when the Zernike of computed image square, true origin is located at the center of image, and pixel value is mapped in the unit circle, i.e. x ²+ y ²≤ 1.Drop on the outer pixel of unit circle and ignore, and have $A_{\mathrm{nm}}^{*}=A_{n.-m}.$ ρ is that initial point arrives that (x, the y) vector distance of pixel, θ are vector ρ and the anticlockwise angle of x axle.V _Nm ^*(ρ θ) is V _Nm(ρ, conjugation θ).V _Nm(ρ, scheduling θ) is as follows:

V _nm(ρ，θ)＝R _nm(ρ)exp(jmθ) (2)

Wherein n is a nonnegative number, m be integer and satisfy simultaneously n-|m| be even number and | m|≤n, R _Nm(ρ) be the radius polynomial expression, the mathematical notation formula is:

$R_{\mathrm{nm}}\left(\mathrm{\&rho;}\right)=\underset{s=0}{\overset{n-\left|m\right|/2}{\mathrm{\&Sigma;}}}{(-1)}^s\frac{(n-s)!}{s!(\frac{n+\left|m\right|}{2}-s)!(\frac{n-\left|m\right|}{2}-s)!}{\mathrm{\&rho;}}^{n-2s}---\left(3\right)$

And R is arranged _{N ,-m}(ρ)=R _Nm(ρ).The amplitude ZMM of Zernike square is | A _Nm|, method for normalizing is | A _Nm| divided by the number of white pixel point in the binary map.ZMM has the robustness to JPEG, noise and rotation, but it is again the description feature of picture shape simultaneously, when picture material is replaced, will cause the bigger variation of ZMM, therefore has half fragile characteristic.The present invention is quantified as 16 with each ZMM, get the highest 4 as watermark.

3. pre-service watermark

If the watermark bit sequence W length that embeds is N, generate 0, the 1 random series R that length is N by secret key k, to watermark sequence W and immediately sequence R carry out xor operation generation watermark information W to be embedded ^*Pretreated purpose is with the watermark information randomization, further improves the accuracy of tampering location.

4. structuring embed watermark

At first seal image is carried out 3 grades of two-dimensional discrete wavelet conversions, obtain 10 subband { LH _i, HL _i, HH _i, LL ₃, i=1-3}, shown in Figure of description 1, LL wherein ₃Subband refers to the subband in the upper left corner, is that the low pass of original seal image is approximate, has concentrated most of energy of seal image signal.Behind 3 grades of wavelet transformations in wavelet coefficient values of LL3 subband and the real image 8 * 8 piece relevant, the piece of LL3 subband 2 * 2 is relevant with 16 * 16 of real images.The present invention is embedded into high 4 forms with 2 * 2 piece of each ZMM in the LL3 subband successively, and shown in Figure of description 2, numeral refers to the sequence number of ZMM among the figure.16 * 16 piece is distorted the value that may change a square in real image.Can determine the position that malice is distorted according to the position of the square that has changed in when authentication.The embedding grammar of each watermark is as follows:

Earlier (a, b) and G_LSB6 (a), wherein (a, b) expression is with low six low six of replacing a of b, hangs down six and a is taken out in G_LSB6 (a) expression for S_LSB6 for two variable S_LSB6 of definition.When the embed watermark bit is ' 1 ', use formula

$F_i^{\&prime;}(u,v)=\left\{\begin{array}{cc}S\_\mathrm{LSB}6(F_i(u,v)-010000b,110000b)& G\_\mathrm{LSB}6\left(F_i(u,v)\right)\&le;010000b\\ S\_\mathrm{LSB}6(F_i(u,v),110000)& \mathrm{otherwise}\end{array}\right.$

When the embed watermark bit is ' 0 ', use formula

$F_i^{\&prime;}(u,v)=\left\{\begin{array}{cc}S\_\mathrm{LSB}6(F_i(u,v)+100000b,010000b)& G\_\mathrm{LSB}6\left(F_i(u,v)\right)\&le;110000b\\ S\_\mathrm{LSB}6(F_i(u,v),010000b)& \mathrm{otherwise}\end{array}\right.$

Wherein, F _i(u v) is a small echo L13 sub-band coefficients, F _i' ' (u v) is that watermark embeds the back coefficient.Carry out wavelet inverse transformation at last, obtain the seal image of embed watermark.

The authentication seal process is as follows:

At first extract the edge and the binaryzation of seal image to be certified, extract the amplitude of 49 Zernike squares on binaryzation edge seal, be called ZMM_A, 49 ZMM_A have represented the feature of image to be certified.Treat authentication image and carry out 3 grades of two-dimensional wavelet transformations, obtain 10 subband { LH _i, HL _i, HH _i, LL ₃, i=1-3} is shown in Figure of description 1.Extract watermark in each piece of 2 * 2 of LL3 subband, the amplitude of 49 of the original seal image of reconstruct Zernike squares is called ZMM_O after the demodulation.

The extracting method of watermark bit is the 6th least significant bit (LSB) of binary representation that obtains the integral part of LL3 sub-band coefficients.

Calculate the quadratic sum of the difference of 49 corresponding ZMM_A and ZMM_O, and judge whether to have surpassed preset threshold 0.125, if it is littler than threshold value, think that then seal is believable, otherwise think that this seal suffered malicious attack, further with the difference of each ZMM_A and ZMM_O square with selected threshold value 0.003 relatively, if greater than threshold value, think that then the embedded location place of this ZMM correspondence has suffered that malice distorts, 16 * 16 of corresponding spatial domain.

Introduce the inventive method below and finish the process that watermark embeds and detects.Here provide the seal image of doing representative with seal 1-4 and done the test result of experiment.Have three aspects: invisibility, the fragility that malice is distorted and tampering location, and to the robustness of normal Flame Image Process such as JPEG compression and noise.

Fig. 3 shows the result who seal 1 image is carried out pre-service and watermark embedding with the inventive method.Wherein Fig. 3 (a) is original seal image, and pretreated seal image is presented among Fig. 3 (b), and the seal image of embed watermark is presented among Fig. 3 (c), and its PSNR is 42.11dB.The seal preprocess method that proposes of the present invention only influences the background of seal as seen from the figure, and the seal that has embedded watermark visually with original seal no significant difference.Satisfy the requirement of invisibility.

Fig. 4 shows the result who seal 2 images is carried out pre-service and watermark embedding with the inventive method.Fig. 4 (a) is original seal image, and pretreated seal image is presented among Fig. 4 (b), and the seal image of embed watermark is presented among Fig. 4 (c), and its PSNR is 42.26dB.The seal preprocess method that proposes of the present invention only influences the background of seal as seen from the figure, and the seal that has embedded watermark visually with original seal no significant difference.Satisfy the requirement of invisibility.

Fig. 5 shows the result who seal 3 images is carried out pre-service and watermark embedding with the inventive method.Fig. 5 (a) is original seal image, and pretreated seal image is presented among Fig. 5 (b), and the seal image of embed watermark is presented among Fig. 5 (c), and its PSNR is 42.21dB.The seal preprocess method that proposes of the present invention only influences the background of seal as seen from the figure, and the seal that has embedded watermark visually with original seal no significant difference.Satisfy the requirement of invisibility.

Fig. 6 shows the result who seal 4 images is carried out pre-service and watermark embedding with the inventive method.Fig. 6 (a) is original seal image, and pretreated seal image is presented among Fig. 6 (b), and the seal image of embed watermark is presented among Fig. 6 (c), and its PSNR is 42.13dB.The seal preprocess method that proposes of the present invention only influences the background of seal as seen from the figure, and the seal that has embedded watermark visually with original seal no significant difference.Satisfy the requirement of invisibility.

Fig. 7 is the authentication result of authentication method when seal 1 image is distorted by malice that the present invention proposes.After malicious attacker replaces to " hundred " word among Fig. 7 (a) " section " word generation Fig. 7 (b), be difficult to tell the true and false of seal with eyes.But thinking that this seal image was carried out artificial distorting when authentication, is incredible, can provide the position of distorting simultaneously, shown in white blocks among Fig. 7 (c).

Fig. 8 is the authentication result of authentication method when seal 2 images are distorted by malice that the present invention proposes.After malicious attacker replaces to " benefit " word among Fig. 8 (a) " dress " word generation Fig. 8 (b), be difficult to tell the true and false of seal with eyes.But thinking that this seal image was carried out artificial distorting when authentication, is incredible, can provide the position of distorting simultaneously, shown in white blocks among Fig. 8 (c).

Fig. 9 is the authentication result of authentication method when seal 3 images are distorted by malice that the present invention proposes.After malicious attacker replaces to " 85 " with " 13 " word among Fig. 9 (a), generates Fig. 9 (b), be difficult to tell the true and false of seal with eyes.But thinking that this seal image was carried out artificial distorting when authentication, is incredible, can provide the position of distorting simultaneously, shown in white blocks among Fig. 9 (c).

Figure 10 is the authentication result of authentication method when seal 4 images are distorted by malice that the present invention proposes.After malicious attacker replaces to " sample " word among Figure 10 (a) " usefulness " word generation Figure 10 (b), be difficult to tell the true and false of seal with eyes.But thinking that this seal image was carried out artificial distorting when authentication, is incredible, can provide the position of distorting simultaneously, shown in white blocks among Figure 10 (c).

The seal verifying method that the present invention proposes can accurately located when malice distorts, for the JPEG compression and add to make an uproar and have robustness preferably.The result is as shown in the table:

	Seal 1	Seal 2	Seal 3	Seal 4
					Q＞＝	30％	40％	40％	30％
S＜＝	4	4	4	4

Wherein Q represents that JPEG quality factor, S represent noise intensity, and image is applied the random noise between [5s, 5s].When the JPEG compression quality factor was not less than value shown in the table, what the seal that compressed through JPEG was considered to be subjected to was normal Flame Image Process.When noise intensity was not higher than value shown in the table, what the seal of crossing through noise pollution was considered to be subjected to was normal Flame Image Process.

Claims (10)

1, a kind of electronic stamp certification method based on characteristics of image is characterized in that method of the present invention is divided into watermark embedding and two processes of authentication seal;

The watermark embed process step is as follows: 1) original seal is carried out pre-service; 2) characteristics of image of the original seal of extraction; 3) pre-service watermark; 4) structuring embed watermark;

The authentication seal process steps is as follows: 1) extract seal image feature to be certified; 2) characteristics of image of the original seal of reconstruct; 3) seal to be certified is authenticated.

2,, it is characterized in that described watermark embed process step 1) carries out pre-service to original seal and adopt following formula according to claims 1 described electronic stamp certification method:

$X^{\&prime;}=\left\{\begin{array}{cc}X& 0\&le;X\&le;T\\ T& T<X\&le;255\end{array}\right.$

Wherein, X represents original seal image pixel value, and X ' is the pixel value of seal image after the pre-service, and T is pretreated threshold value, gets 245.

3, according to claims 1 described electronic stamp certification method, it is characterized in that described watermark embed process step 2) the characteristics of image method of extracting original seal is as follows: utilizes the edge extracting algorithm to extract the edge of seal figure and carries out binaryzation, on binary image, calculate the amplitude of 49 Zernike squares, be designated as ZMM, and it is normalized to decimal between the 0-1; Each ZMM is quantified as 16, get the highest 4 as watermark.

4, electronic stamp certification method according to claim 1, it is characterized in that described watermark embed process step 3) pretreated water impression method is as follows: the watermark bit sequence W length of establishing embedding is N, generate 0, the 1 random series R that length is N by secret key k, to watermark sequence W and immediately sequence R carry out xor operation generation watermark information W to be embedded ^*

5, electronic stamp certification method according to claim 1 is characterized in that described watermark embed process step 4) structuring embed watermark method is as follows: at first seal image is carried out 3 grades of two-dimensional discrete wavelet conversions, obtain 10 subband { LH _i, HL _i, HH _i, LL ₃, i=1-3}, wherein LL ₃Subband is that the low pass of original seal image is approximate, has concentrated most of energy of seal image signal, LL behind 3 grades of wavelet transformations ₃8 * 8 piece is relevant in wavelet coefficient values of subband and the real image, LL ₃The piece of subband 2 * 2 is relevant with 16 * 16 of real images; Successively high 4 forms with 2 * 2 piece of each ZMM are embedded into LL then ₃In the subband, 16 * 16 piece is distorted the value that may change a square in real image, can determine the position that malice is distorted according to the position of the square that has changed in when authentication; Carry out wavelet inverse transformation at last, obtain the seal image of embed watermark.

6, electronic stamp certification method according to claim 1, it is characterized in that described authentication seal process steps 1) to extract seal image characterization method to be certified as follows: the edge and the binaryzation of getting seal image to be certified, on binaryzation edge seal, extract the amplitude of 49 Zernike squares, be called ZMM_A, 49 ZMM_A have represented the feature of image to be certified.

7, electronic stamp certification method according to claim 1 is characterized in that described authentication seal process steps 2) the characteristics of image method of the original seal of reconstruct is as follows: at first treats authentication image and carries out 3 grades of two-dimensional wavelet transformations, obtain 10 subband { LH _i, HL _i, HH _i, LL ₃, i=1-3}; In each piece of 2 * 2 of LL3 subband, extract watermark then, obtain watermark information W ', by identical secret key k in the watermark pretreatment when embedding, generate random series R, to watermark sequence W ' and immediately sequence R carry out xor operation and generate watermark information W "; " amplitude of 49 Zernike squares of the original seal image of reconstruct is called ZMM_O by W.

8, electronic stamp certification method according to claim 1, it is characterized in that described authentication seal process steps 3) that seal to be certified is carried out authentication method is as follows: the quadratic sum of calculating the difference of 49 corresponding ZMM_A and ZMM_O, and judge whether to have surpassed preset threshold 0.125, if it is littler than threshold value, think that then seal is believable, otherwise think that this seal suffered malicious attack, further with the difference of each ZMM_A and ZMM_O square with selected threshold value 0.003 relatively, if greater than threshold value, think that then the embedded location place of this ZMM correspondence has suffered that malice distorts, 16 * 16 of corresponding spatial domain.

9, electronic stamp certification method according to claim 5 is characterized in that the embedding grammar of each watermark in the described watermark embed process step 4) is as follows:

Define two variable S_LSB6 (a, b) and G_LSB6 (a), wherein S_LSB6 (a, b) expression is with low six low six of replacing a of b, and G_LSB6 (a) expression is taken out low six of a, when the embed watermark bit be ' 1 ', the use formula:

$F_i^{\&prime;}(u,v)=\left\{\begin{array}{cc}S\_\mathrm{LSB}6(F_i(u,v)-010000b,110000b)& G\_\mathrm{LSB}6\left(F_i(u,v)\right)\&le;010000b\\ S\_\mathrm{LSB}6(F_i(u,v),110000)& \mathrm{otherwise}\end{array}\right.$

When the embed watermark bit is ' 0 ', use formula

$F_i^{\&prime;}(u,v)=\left\{\begin{array}{cc}S\_\mathrm{LSB}6(F_i(u,v)+100000b,010000b)& G\_\mathrm{LSB}6\left(F_i(u,v)\right)\&le;110000b\\ S\_\mathrm{LSB}6(F_i(u,v),010000b)& \mathrm{otherwise}\end{array}\right.$

Wherein, F _i(u v) is small echo LL ₃Sub-band coefficients, F _i' (u v) is that watermark embeds the back coefficient.

10, electronic stamp certification method according to claim 7 is characterized in that described authentication seal process steps 2) in the extracting method of watermark bit be to obtain LL ₃The 6th least significant bit (LSB) of the binary representation of the integral part of sub-band coefficients.

Images