CN103345725A - Volume data watermarking method based on three-dimensional DWT-DFT and chaos scrambling - Google Patents

Abstract

The invention discloses a volume data digital watermarking technique based on three-dimensional DWT-DFT and chaos scrambling and belongs to the field of multi-media signal processing. The technique comprises the steps of carrying out chaos scrambling on a watermark by means of the Logistic Map nature; extracting a feature vector by carrying out three-dimensional DWT-DFT on original volume data to embed the watermark, correlating the feature vector with the chaos scrambled watermark to obtain a two-valued logic sequence, and saving the two-valued sequence in a third party; extracting a feature vector by carrying out three-dimensional DWT-DFT on volume data to be tested and extracting the watermark by correlating the feature vector with the two-valued sequence saved in the third party; restoring the watermark by means of the Logistic Map nature. According to the volume data digital watermarking technique based on the three-dimensional DWT-DFT and the chaos scrambling, robustness is good, and the content of the original volume data is not changed by the embedding of the watermark.

Description

Volume data water mark method based on three-dimensional DWT-DFT and chaos scramble

Technical field

The present invention relates to a kind of digital watermark technology of the volume data characteristics of image based on 3 D wavelet transformation (DWT), three-dimensional Fourier transform (DFT) and chaos scramble, is a kind of multimedia data protection method, belongs to field of multimedia signal processing.

Technical background

Utilize the internet can realize distance medical diagnosis; but when passing through internet transmission patient's medical imaging; may reveal patient's personal information; how to protect the patient individual privacy, make that the personal information in patient's the medical imagings such as CT, MRI, patient's data such as electronic health record are not revealed, this problem is along with the universal of internet becomes serious day by day.Medical image digital watermark technology (Medical Image Watermarking is called for short MIW) can address this problem effectively.

Digital watermark technology is the copyright protection for the Digital Media on the internet at first; utilize the characteristics such as invisibility, robustness of digital watermarking now; can be hidden in patient's personal information in its medical image, to guarantee its safe transmission on the internet.The appearance of medical image digital watermarking, make distance medical diagnosis, when the required relevant patient data of remote operation transmits on the internet, can effectively protect patient's privacy, avoid patient's data to be distorted.

At present, in the research field of digital watermarking to how in volume data the Study on Problems of embed watermark less because medical image (CT, MRI etc.) major part is volume data, these volume datas are not allow to revise its content in principle.In addition, image compression standard JPEG 2000 of future generation is based on wavelet transformation.Therefore, to based on three-dimensional DWT-DFT, the work of embed watermark has big meaning in volume data, and the watermark that requires to embed has stronger robustness, realizes that difficulty is bigger, has not yet to see report, still belongs to blank.

Summary of the invention

The objective of the invention is to propose a kind of based on three-dimensional DWT-DFT and Logistic Map chaos scramble; realization can be resisted volume data digital watermarking embedding and the extracting method that geometric attack can be resisted conventional attack again; it has stronger robustness; and the embedding of watermark does not influence the initial body data; be a kind of zero watermarking project, thus better protect three-dimensional data.

Principle of the present invention is: at first utilize the watermark information of Logistic Map to carry out pre-service, again medical volume data is carried out overall three-dimensional DWT-DFT conversion, in three-dimensional DWT-DFT conversion coefficient, extract the proper vector of a resist geometric attacks, and pretreated watermark is associated with this proper vector, utilize the robustness of proper vector to realize anti-geometry and the conventional attack of digital watermarking.

To achieve these goals, the present invention is performed such: use Logistic Map and produce chaos sequence (i j) carries out chaos scramble and reduction, improves the security of watermark information to the watermarking images W of two-value.By volume data is carried out overall 3 D wavelet transformation, obtain " approximation coefficient " and " detail coefficients ", the wavelet transformation of this similar two dimensional image, " approximation coefficient " represents the low frequency characteristic of volume data, reflection be the main exterior contour of volume data; " detail coefficients " represents the high frequency characteristics of volume data, reflection be the detail of the high frequency of volume data.The resist geometric attacks ability of wavelet transformation itself is relatively poor, therefore, we carry out 3 D wavelet transformation (DWT) to volume data earlier, and then to the reflection low frequency characteristic " approximation coefficient " carry out overall Fourier transform (DFT), in three-dimensional DFT coefficient, extract the proper vector of a resist geometric attacks of volume data, and the Hash function in digital watermark and the cryptography and " third party " concept are combined, realized the digital watermark technology based on three-dimensional DWT-DFT resist geometric attacks.The method applied in the present invention comprises the chaos scramble of watermark, embedding, extraction and reduction four major parts of watermark, and first is the chaos scramble of watermark, comprising: (1) produces chaos sequence X (j) by Logistic Map; (2) according to X (j) scramble is carried out in watermark, obtain the chaos scramble watermark BW (i, j); Second portion is the embedding of watermark, comprising: (3) by volume data is carried out 3 D wavelet transformation, the pairing approximation coefficient carries out overall DFT conversion then, obtains the proper vector V (j) of a resist geometric attacks of volume data; (4) according to the watermark BW of chaos scramble (i, j) and the proper vector V (j) of the volume data of extracting, by the Hash functional operation, generate a two-valued function key sequence Key (i, j), (j) there is the third party in i with two-valued function sequence Key then; Third part is the extraction of watermark, comprising: (5) obtain volume data to be measured resist geometric attacks proper vector V ' (j); (6) utilize be present in third-party two-valued function key sequence Key (i, j) and the proper vector V ' of volume data to be measured (j), extract watermark BW ' (i, j); The 4th part is the reduction of watermark, comprising: (7) are used Logistic Map and are obtained identical chaos sequence X (j); (8) by X (j) watermark is reduced.

Now be elaborated as follows to method of the present invention:

At first select a significant bianry image as the watermark that will embed medical volume data, be designated as W={w (i, j) | w (i, j)=0,1; 1≤i≤M1,1≤j≤M2}; Simultaneously, choose a MRI volume data carrying among the Matlab as the primitive medicine volume data, be expressed as: F={f (i, j, k) | f (i, j, k) ∈ R; 1≤i≤M, 1≤j≤N, 1≤k≤P}.Wherein, w (i, j) and f (i, j k) represent the grey scale pixel value of watermark and voxel (Voxel) data value of primitive medicine volume data respectively, the grey scale pixel value in this similar two dimensional image, convenient for the purpose of, establish M1=M2, M=N.

First: the chaos scramble of watermark

1) generates chaos sequence by Logistic Map;

By initial value x ₀Generate chaos sequence X (j) by Logistic Map chaos system.

2) obtain the watermark of chaos scramble;

At first, the value among the chaos sequence X (j) is sorted according to ascending order, changes according to each value ordering front-back direction among the X (j) then scramble is carried out in watermark locations of pixels space, obtain the chaos scramble watermark BW (i, j).

Second portion: the embedding of watermark

3) by the initial body data are carried out 3 D wavelet transformation, " approximation coefficient " to wavelet transformation carries out overall DFT conversion more then, in the Low Medium Frequency coefficient of DFT, obtains the proper vector V (j) of a resist geometric attacks of this volume data;

(i, j k) carry out overall three-dimensional DWT conversion, obtain ll channel coefficient FA to initial body data F earlier _L, again to ll channel FA _LCarry out overall three-dimensional DFT conversion, obtain matrix of coefficients FF (i, j, k), L Low Medium Frequency coefficient value before taking out, and pass through FF (i, j, k) coefficient carries out the proper vector V (j) that symbolic operation obtains this volume data, when coefficient value be on the occasion of or during null value we with " 1 " expression, with " 0 " expression, process prescription is as follows during for negative value:

FA _L(i,j,k)=DWT3(F(i,j,k))

FF(i,j,k)=DFT3(FA _L(i,j,k))

V(j)=Sign(FF(i,j,k))

4) according to the watermark BW of chaos scramble (i j) and the proper vector V (j) of the volume data of extracting, utilizes Hash function character, generate a two-valued function sequence Key (i, j);

Key(i,j)=V(j)⊕BW(i,j)

(i is that (i, j), the Hash function commonly used by cryptography generates by the proper vector V (j) of volume data and the watermark BW of chaos scramble j) to Key.(i j), need use when extracting watermark afterwards to preserve Key.By (i j) applies for to the third party as key, to obtain entitlement and the right to use of medical volume data, reaches the purpose of copyright protection with Key.

Third part: the extraction of watermark

5) the proper vector V ' that obtains volume data to be measured (j);

If volume data to be measured is that (i, j k), obtain ll channel coefficient FA ' through 3 D wavelet transformation (DWT) to F ' _L, again to ll channel FA ' _LCarry out overall three-dimensional DFT conversion, obtain the method for matrix of coefficients FF ' (i, j k), press above-mentioned steps 3), try to achieve the proper vector V ' of volume data to be measured (j), process prescription is as follows:

FA’ _L(i,j,k)=DWT3(F’(i,j,k))

FF’(i,j,k)=DFT3(FA’ _L(i,j,k))

V’(j)=Sign(FF’(i,j,k))

6) in volume data to be measured, extract watermark BW ' (i, j);

According to the logic key sequence Key that generates when the embed watermark (i, j) and the proper vector V ' of volume data to be measured (j), the watermark BW ' that utilizes Hash function character to extract to contain in the volume data to be measured (i, j).

BW’(i,j)=Key(i,j)⊕V’(j)

The 4th part: the reduction of watermark

7) generate chaos sequence by Logistic Map;

By generating identical chaos sequence X (j) with initial value x0 that above step 1) is identical by Logistic Map chaos system;

8) watermark of reduction extraction;

At first with ascending ordering of value among the chaos sequence X (j), then according to each value ordering front-back direction among the X (j) change to watermark locations of pixels space reduce the watermark W ' that obtains reducing (i, j).

Again according to W (i, j) and W ' (i, degree of correlation j) is differentiated the entitlement of volume data to be measured.

The present invention has compared following advantage with existing medical science digital watermark:

At first, the present invention is based on the digital watermark technology of three-dimensional DWT and three-dimensional DFT, DWT is the core of Image Compression JPEG2000 of future generation, DFT is in frequency field, can find the proper vector of resist geometric attacks therein, experimental data by the back confirms that this water mark method not only has stronger anti-conventional attack ability, and stronger resist geometric attacks ability is arranged; Secondly, embedding be through the watermark of Logistic Map chaos scramble, watermark information is become disorderly and unsystematic, improved the security of watermark information; At last, the embedding of watermark does not influence the content of initial body data, is a kind of zero digital watermark, better protection medical volume data.This characteristic especially has very high practical value at aspects such as medical image processing, and usable range is wide, and can realize embedding and the extraction of many watermarks and big watermark.

Below we from the explanation of theoretical foundation and experimental data:

1) 3 d-dem wavelet transformation (DWT)

One deck decomposable process of 3 D wavelet transformation as shown in Figure 1, L among Fig. 1, H represent low-frequency component and the radio-frequency component of medical volume data through obtaining after low frequency and the High frequency filter respectively, similar with the wavelet transformation of two dimensional image, medical volume data is broken down into " approximation coefficient " LLL who represents the volume data low frequency characteristic through behind the 3 D wavelet transformation ₁(low frequency 3-d subband) and represent " detail coefficients " (high frequency 3-d subband) of the high-frequency information of this volume data, the ground floor of the three-dimensional DWT of subscript " 1 " expression decomposes; The example of the 3 D wavelet transformation of one individual data items (two-layer) is seen Fig. 2 to Fig. 4, and Fig. 2 is a section of volume data, and Fig. 3 is the three-dimensional imaging of volume data, and Fig. 4 is the 3 D wavelet transformation (two-layer) of volume data.

2) 3 d-dem Fourier transform

Function f (3 d-dem Fourier direct transform formula z) is for x, y:

$F(u,v,w)=\underset{x=0}{\overset{M-1}{\mathrm{\Σ}}}\underset{y=0}{\overset{N-1}{\mathrm{\Σ}}}\underset{z=0}{\overset{P-1}{\mathrm{\Σ}}}f(x,y,z)\·e^{-j2\mathrm{\πxu}/M}{e}^{-j2\mathrm{\πyv}/N}{e}^{-j2\mathrm{\πzw}/P}$

u＝0,1,...,M-1;v＝0,1,...,N-1;w＝0,1,...,P-1;

The 3 d-dem Fourier inversion formula is:

$f(x,y,z)=\frac{1}{\mathrm{MNP}}\underset{u=0}{\overset{M-1}{\mathrm{\Σ}}}\underset{v=0}{\overset{N-1}{\mathrm{\Σ}}}\underset{w=0}{\overset{P-1}{\mathrm{\Σ}}}F(u,v,w)e^{j2\mathrm{\πxu}/M}{e}^{j2\mathrm{\πyv}/N}{e}^{j2\mathrm{\πzw}/P}$

x＝0,1,...,M-1;y＝0,1,...,N-1;z＝0,1,...,P-1;

In Fourier's positive inverse transform, (x, y z) are three dimensions territory function to f, and (u, v w) are corresponding frequency-domain function to F.

3）Logistic?Map

Chaos is a kind of random motion that seems to be, and refers to the similar process at random that occurs in deterministic system.Therefore, its initial value and parameter arranged, we just can generate this chaos system.Logistic Map is foremost a kind of chaos system, and it is the Nonlinear Mapping that is given by the following formula:

x _k+1＝μx _k(1-x _k)

Wherein, 0≤μ≤4 are growth parameter, x _k∈ (0,1) is system variable, and k is iterations.The research work of Chaos dynamic system points out, when growth parameter 3.569945≤μ≤4, Logistic Map works in chaos state.Can see that initial value has a slight difference will cause the significant difference of chaos sequence.Therefore, above sequence is a desirable key sequence.Set μ=4 herein, chaos sequence is by different initial value x ₀Produce.

4) extraction of the proper vector of the resist geometric attacks of volume data

The main cause of present most of watermarking algorithm resist geometric attacks ability is: people are embedded in digital watermarking in voxel or the conversion coefficient, and the slight geometric transformation of volume data usually causes the bigger variation of voxel data value or transform coefficient values.The watermark that is embedded in like this in the volume data is just attacked easily.If can find the proper vector of a reflection medical volume data geometrical feature, and when little geometric transformation takes place in volume data, tangible sudden change can not take place in this proper vector value, and the watermark that will embed and this proper vector are associated, just can solve the robustness problem of watermark preferably.The ability of wavelet transformation resistance geometric attack is relatively poor, and data are found 3 D wavelet transformation and the three-dimensional Fourier transform of volume data are combined by experiment, can find the proper vector of a resist geometric attacks.(realize by geometric transformation is carried out in each section) that when an individual data items is carried out common geometric transformation some variations may take place the size of three-dimensional DFT Low Medium Frequency coefficient value (referring to real part, imaginary part coefficient), but its coefficient symbols remains unchanged substantially.According to this rule of finding, we carry out 3 D wavelet transformation (selecting one deck here for use) to volume data earlier, then its approximation coefficient are carried out overall three-dimensional DFT conversion again, and we illustrate by some experimental datas of table 1.The former figure that is used as test in the table 1 is Fig. 5, it is a section (getting the tenth) of a MRI volume data carrying among the matlab, " the 1st row " demonstration is volume data type under attack in the table 1, this sectioning image that is subjected to behind the conventional attack is seen Fig. 6 to Fig. 8, and Fig. 9 to Figure 12 is seen in the three-dimensional imaging of conventional attack correspondence; The sectioning image that is subjected to behind the geometric attack is seen Figure 13 to Figure 16, and Figure 17 to Figure 20 is seen in its corresponding three-dimensional imaging." the 2nd row " of table 1 are FF (1,1,1)-12 Low Medium Frequency coefficients of FF (1,2,3) (plural number is calculated here and done two coefficients) of getting in the three-dimensional DWT-DFT matrix of coefficients to " the 7th row ".For conventional attack, these Low Medium Frequency coefficient values remain unchanged (first row represent direct current component, size variation is big slightly) and the three-dimensional DWT-DFT coefficient value approximately equal of initial body data substantially; For geometric attack, most of coefficient has bigger variation, but can find from table 1, and the size of most of three-dimensional DWT-DFT Low Medium Frequency coefficient has taken place to change but its symbol does not have to change substantially.We represent positive three-dimensional DWT-DFT coefficient (containing value is zero coefficient), " 0 " expression of negative coefficient with " 1 ".So for the initial body data, FF (1,1,1)-FF (1 in the three-dimensional DWT-DFT matrix of coefficients, 2,3) the coefficient symbols sequence of coefficient correspondence is: " 110000110011 ", specifically see Table 1 " the 8th row ", observe these row and can find, no matter conventional attack still is that the symbol sebolic addressing of this symbol sebolic addressing of geometric attack and initial body data keeps similar, all bigger with the normalized correlation coefficient of initial body data corresponding symbol sequence, be 1, see Table 1 " the 9th row ").

In order to prove that further the proper vector of extracting as stated above is a key character of this volume data, we are again different tested objects (seeing Figure 21 to Figure 26), carry out three-dimensional DWT-DFT conversion according to the method described above, obtain corresponding conversion coefficient FF (1,1,1)-FF (1,2,3), and obtain related coefficient with the symbol sebolic addressing of initial body data, the result is as shown in table 2.

By data in the table 2 as can be seen, the symbol sebolic addressing between the different volume datas differs bigger, and the degree of correlation is less.

Table 1 three-dimensional DWT-DFT low frequency part coefficient and be subjected to different the attack after changing value

* DWT-DFT conversion coefficient unit: 1.0e+005

The three-dimensional DWT-DFT low frequency coefficient of the different volume datas of table 2 and the degree of correlation of symbol sebolic addressing thereof

The unit of Figure 25 in the table: 1.0e+008; The unit of other figure: 1.0e+005;

In sum, we pass through the analysis to the three-dimensional DWT-DFT coefficient of volume data, utilize the symbol sebolic addressing of Low Medium Frequency coefficient to obtain a method of resisting the proper vector of geometric attack, utilize this proper vector and Hash function character, " third party " concept to realize the method for embed watermark in volume data.Experimental result proves, this method has realized the blind extraction of zero watermark embedding and watermark, better must protect the primitive medicine volume data.

Description of drawings

Fig. 1 is 3 D wavelet transformation synoptic diagram (one deck).

Fig. 2 is a section of initial body data.

Fig. 3 is the three-dimensional imaging of initial body data correspondence.

Fig. 4 is result's demonstration of the initial body data being carried out 3 D wavelet transformation (two-layer).

Fig. 5 is a section (acquiescence is the 10th section) of initial body data.

Fig. 6 is that intensity is the sectioning image after 3% Gaussian noise is disturbed.

Fig. 7 is through the sectioning image after the JPEG compression (compression quality is 4%).

Fig. 8 is through the sectioning image behind the medium filtering (filtering parameter is [3x3]).

Fig. 9 is the three-dimensional imaging of initial body data correspondence.

Figure 10 is that intensity is 3% the corresponding three-dimensional imaging in Gaussian noise interference back.

Figure 11 is through the corresponding three-dimensional imaging in JPEG compression (compression quality is 4%) back.

Figure 12 is through three-dimensional imaging corresponding behind the medium filtering (filtering parameter is [3x3]).

Figure 13 is the sectioning image of up time rotation 20 degree.

Figure 14 is the sectioning image of 0.5 times of convergent-divergent.

Figure 15 is that vertical direction moves down 8% sectioning image.

Figure 16 is that Z-direction is sheared first sectioning image after 8%.

Figure 17 is the volume data three-dimensional imaging of up time rotation 20 degree.

Figure 18 is the volume data three-dimensional imaging of 0.5 times of convergent-divergent.

Figure 19 is that vertical direction moves down 8% volume data three-dimensional imaging.

Figure 20 is that Z-direction is sheared 8% volume data three-dimensional imaging.

Figure 21 is the three-dimensional imaging of volume data MRI_1.

Figure 22 is the three-dimensional imaging of volume data MRI_2.

Figure 23 is the three-dimensional imaging of volume data MRI_3.

Figure 24 is the three-dimensional imaging of volume data Teddy bear.

Figure 25 is the three-dimensional imaging of volume data Tooth.

Figure 26 is the three-dimensional imaging of volume data Liver.

Figure 27 is original watermark.

Figure 28 is through the watermark behind the Logistic Map chaos scramble.

Figure 29 is the watermark section that does not add when disturbing.

Figure 30 is the three-dimensional reconstruction figure that does not add when disturbing.

Figure 31 does not add the watermark of extracting when disturbing.

Figure 32 is the sectioning image (Gaussian noise intensity 5%) after Gaussian noise is disturbed.

Figure 33 is the three-dimensional reconstruction figure (Gaussian noise intensity 5%) after Gaussian noise is disturbed.

Figure 34 is the watermark (Gaussian noise intensity 5%) that Gaussian noise disturbs the back to extract.

Figure 35 is the sectioning image (the compression quality parameter is 2%) after the JPEG compression.

Figure 36 is the volume data three-dimensional imaging (the compression quality parameter is 2%) after the JPEG compression.

Figure 37 is the watermark (the compression quality parameter is 2%) that extract JPEG compression back.

Figure 38 is the sectioning image (filtering parameter is [5x5], and filter times is 10 times) behind the medium filtering.

Figure 39 is the volume data three-dimensional imaging (filtering parameter is [5x5], and filter times is 10 times) behind the medium filtering.

Figure 40 is the watermark (filtering parameter is [5x5], and filter times is 10 times) of extracting behind the medium filtering.

Figure 41 is the sectioning image behind up time rotation 10 degree.

Figure 42 is the three-dimensional imaging of up time rotation 10 degree back volume datas.

Figure 43 is the watermark that extract up time rotation 10 degree backs.

Figure 44 is that zoom factor is 0.5 sectioning image.

Figure 45 is that zoom factor is 0.5 three-dimensional imaging.

Figure 46 is that zoom factor is the watermark of extracting in 0.5 o'clock.

Figure 47 is the sectioning image after the horizontal left 10%.

Figure 48 is the three-dimensional imaging of horizontal left 10% back volume data correspondence.

Figure 49 is the watermark that extract horizontal left 10% back.

Figure 50 is after Z-direction shears 20%, first sectioning image of volume data.

Figure 51 is after Z-direction shears 20%, the three-dimensional imaging of volume data.

Figure 52 is after Z-direction shears 20%, the watermark of extraction.

Embodiment

The invention will be further described below in conjunction with accompanying drawing, selects a significant bianry image as original watermark, is designated as: and W={w (i, j) | w (i, j)=0,1; 1≤i≤M1,1≤j≤M2} sees Figure 27, the size of watermark here is 32 * 32.See Figure 28 by the watermark behind the Logistic Map chaos scramble, can see obviously that very big variation has taken place in watermark, security improves.Fig. 5 is seen in a section of primitive medicine volume data, is to take from the nuclear magnetic resonance 3-D view volume data (MRI.mat) that carries among the matlab, and the size of volume data is 128x128x27, sees Fig. 9.The initial body data be expressed as F (i, j, k), 1≤i wherein, j≤128; 1≤k≤27, corresponding three-dimensional DWT-DFT matrix of coefficients be FF (i, j, k), 1≤i wherein, j≤128; 1≤k≤27.Consider robustness and disposable embed watermark capacity we get 32 coefficients.By watermarking algorithm detect W ' (i, j) after, we have judged whether that by calculating normalized correlation coefficient NC watermark embeds.

Do not add the watermarking images (acquiescence is selected the tenth section, and test is made up of 27 sections altogether with volume data) when disturbing here.

Figure 29 is the watermark section that does not add when disturbing;

Figure 30 is the volume data three-dimensional imaging that does not add when disturbing;

Figure 31 does not add the watermark of extracting when disturbing, and can see NC=1.00, can accurately must extract watermark.

Below we judge anti-conventional attack ability and the resist geometric attacks ability of this digital watermark method by concrete experiment.

Test the ability of the anti-conventional attack of this watermarking algorithm earlier.

(1) adds Gaussian noise

Use imnoise () function in watermarking images, to add Gaussian noise.

Table 3 is the anti-Gaussian noise interference experiment of watermark data.Data can be seen from table, when noise intensity up to 25% the time, the PSNR of watermark volume data is down to 0.10dB, at this moment the watermark related coefficient NC=0.80 of Ti Quing can accurately extract watermark.Therefore, this watermarking algorithm has good anti-Gaussian noise interference performance.

Figure 32 is that Gaussian noise intensity is 5% o'clock watermark section, and is visually very fuzzy;

Figure 33 is corresponding volume data three-dimensional imaging, and volume data is visually very fuzzy, and the PSNR value is lower, is 6.00dB;

Figure 34 is the watermark of extracting, and NC=0.93 can accurately must extract watermark.

The anti-Gaussian noise interference experiment of table 3 watermark data

Noise intensity (%)	3	5	10	15	20	25
							PSNR(dB)	8.03	6.00	3.31	1.82	0.82	0.10
NC	0.93	0.93	0.92	0.87	0.81	0.80

(2) the JPEG compression is handled

Adopt image compression quality percentage as parameter volume data to be carried out the JPEG compression.

Table 4 is the anti-JPEG compression experiment of watermark data.When compression quality only was 2%, compression quality was lower, PSNR=16.57dB, and NC=0.88 still can accurately extract watermark.

Figure 35 is that compression quality is 2% sectioning image, and blocking artifact has appearred in this figure;

Figure 36 is that compression quality is the three-dimensional imaging of 2% volume data correspondence, and three-dimensional blocking artifact has appearred in this figure;

Figure 37 is the watermark of extracting, and NC=0.88 can accurately extract watermark.

The anti-JPEG compression experiment of table 4 watermark data

Compression quality (%)	2	4	8	10	20	40	60
								PSNR(dB)	16.57	17.82	20.21	21.20	23.10	25.06	26.61
NC	0.88	0.76	0.93	0.87	0.87	0.93	0.93

(3) medium filtering is handled

Table 5 is the anti-medium filtering experimental datas of watermark volume data, from table data as can be seen, when the medium filtering parameter is [5x5], the filtering multiplicity is 10 o'clock, PSNR=18.69dB is worth lowlyer, but still can accurately extract watermark, NC=0.92.

Figure 38 is that the medium filtering parameter is [5x5], and the filtering multiplicity is 10 sectioning image, and bluring has appearred in image;

Figure 39 is corresponding volume data three-dimensional imaging, and PSNR=18.69dB be worth lowlyer, and the ear profile of three-dimensional imaging at this moment etc. are partly not too clearly demarcated;

Figure 40 is the watermark of extracting, and NC=0.92 can accurately extract watermark.

The anti-medium filtering experimental data of table 5 watermark

Watermark resist geometric attacks ability

(1) rotational transform

Table 6 is the anti-rotational transform experimental datas of watermark, can see from table when watermark volume data up time rotation 15 is spent, and still can extract watermark, at this moment NC=0.74.And the volume data watermarking algorithm that Y.H.WU provides in paper in 2000, when rotation only is 1.50 when spending, normalized correlation coefficient is just lower, and NC=0.24 can't extract watermark.

Figure 41 is the sectioning image of 10 ° of up time rotations;

Figure 42 is corresponding volume data three-dimensional imaging, PSNR=13.97dB, and signal to noise ratio (S/N ratio) is lower;

Figure 43 is the watermark of extracting, and NC=0.78 can accurately extract watermark.

Experimental data is attacked in the anti-rotation of table 6 watermark

The up time rotation	2 degree	5 degree	7 degree	10 degree	13 degree	15 degree
							PSNR(dB)	20.97	16.54	15.17	13.97	13.28	12.98
NC	1.00	0.85	0.85	0.78	0.73	0.74

(2) scale transformation

Table 7 is the nonshrink attack experimental datas of putting of watermark volume data, as can be seen from Table 7 when the zoom factor of watermarking images little to 0.5 the time, related coefficient NC=0.81 still can extract watermark.So this water mark method has the stronger nonshrink attacking ability of putting.

Figure 44 is the watermark sectioning image (zoom factor is 0.5) after convergent-divergent is attacked;

Figure 45 is the three-dimensional imaging (zoom factor is 0.5) that convergent-divergent is attacked back volume data correspondence;

Figure 46 is that convergent-divergent is attacked the watermark that extract the back, and NC=0.81 can accurately extract watermark.

The nonshrink attack experimental data of putting of table 7 watermark

Zoom factor	0.5	0.7	0.9	1.0	1.2	2.0	4.0
								NC	0.81	0.88	0.93	1.00	1.00	0.79	0.71

(3) translation transformation

Table 8 is the anti-translation transformation experimental datas of watermark, learns that from table volume data PSNR=9.27dB has NC=0.73, still can accurately must extract watermark, so this water mark method has stronger anti-translation transformation ability when horizontal left 12%.

Figure 47 is the watermark sectioning image of horizontal left 10%;

Figure 48 is corresponding volume data three-dimensional imaging, PSNR=9.80dB, and signal to noise ratio (S/N ratio) is lower;

Figure 49 is the watermark of extracting, and NC=0.73 can accurately must extract watermark.

The anti-translation transformation experimental data of table 8 watermark (horizontal left)

Horizontal left (%)	2	4	6	8	10	12
							PSNR(dB)	13.02	11.38	10.90	10.21	9.80	9.27
NC	1.00	1.00	0.95	0.73	0.73	0.73

(4) shearing attack

Table 9 is the anti-shearing attack experimental datas of watermark, and data can be seen from table, when shearing from Z-direction up to 24% the time, still can extract watermark, at this moment NC=0.87.Illustrate that this water mark method has stronger anti-shearing attacking ability.

Figure 50 is after shearing 20% by Z-direction, first sectioning image of volume data;

Figure 51 shears the corresponding volume data three-dimensional imaging in 20% back by Z-direction, can see that the effect of shearing attack is obvious, and the former relatively figure in top has cut greatly;

Figure 52 is the watermark of extracting, and NC=0.87 can accurately extract watermark.

The anti-shearing attack experimental data of table 9 watermark (shearing by Z-direction)

Shearing ratio (%)	4	6	8	10	14	16	18	20	22	24
											NC	0.93	0.87	0.87	0.87	0.87	0.87	0.87	0.87	0.87	0.87

By above description of test, this water mark method has stronger anti-conventional attack and geometric attack ability, and the embedding of watermark do not influence the initial body data, is a kind of zero watermarking project.

Claims (1)

1. based on the volume data water mark method of three-dimensional DWT-DFT and chaos scramble, it is characterized in that: based on three-dimensional DWT-DFT conversion, obtain the proper vector of the resist geometric attacks of medical volume data, and chaos Chaotic Technology and digital watermark combined, resist geometric attacks and the conventional attack of medical volume data digital watermarking have been realized, this volume data digital watermarking implementation method is divided into four parts, amounts to eight steps:

First is the chaos scramble of watermark: utilize Logistic Map character to produce chaos sequence scramble carried out in watermark, obtain the chaos scramble watermark BW (i, j);

1) by logic initial value x ₀Generate chaos sequence X (j) by Logistic Map;

2) value among the chaos sequence X (j) is arranged according to from small to large order, according to each value ordering front-back direction variation among the X (j) scramble is carried out in the locus of watermark pixel again, obtain the chaos scramble watermark BW (i, j);

Second portion is the embedding of watermark: by the embedding operation to watermark, obtain corresponding two-valued function sequence Key (i, j);

3) the initial body data are carried out 3 D wavelet transformation, the pairing approximation coefficient carries out overall Fourier transform again, in the DWT-DFT conversion coefficient, obtains the proper vector V (j) of a resist geometric attacks of this volume data according to the symbol sebolic addressing of Low Medium Frequency coefficient;

4) utilize Hash function character and chaos scramble watermark BW (i, j), obtain a two-valued function sequence Key (i, j), $\mathrm{Key}(i,j)=V\left(j\right)\⊕\mathrm{BW}(i,j);$

(i j), will use when extracting watermark below, by (i j) applies for to the third party as key, to obtain the entitlement to the primitive medicine volume data Key to preserve Key;

Third part is the extraction of watermark: by two-valued function sequence Key (i, j) and the proper vector V ' of the resist geometric attacks of volume data to be measured (j), extract watermark BW ' (i, j);

5) volume data to be measured is carried out 3 D wavelet transformation, the pairing approximation coefficient carries out overall DFT conversion again.In conversion coefficient, obtain according to the symbol sebolic addressing of Low Medium Frequency coefficient volume data to be measured a resist geometric attacks proper vector V ' (j);

6) utilize Hash function character and be present in third-party Key (i j), extracts watermark, ${\mathrm{BW}}^{,}(i,j)=\mathrm{Key}(i,j)\⊕V^{,}\left(j\right);$

The 4th part is the reduction of watermark: utilize Logistic Map character to obtain chaos sequence X (j), the reduction watermark;

7) by logic initial value x ₀Generate chaos sequence X (j);

8) to ascending ordering of value among the chaos sequence X (j), change according to each value ordering front-back direction among the X (j), the locus of the watermark pixel extracted is reduced, the watermark W ' that obtains reducing (i, j);

With W (i, j) and W ' (i j) carries out normalized correlation coefficient and calculates, and determines the entitlement of medical volume data.

Images