CN1937769B - Water print method and device for digital video signal - Google Patents
Water print method and device for digital video signal Download PDFInfo
- Publication number
- CN1937769B CN1937769B CN 200610113607 CN200610113607A CN1937769B CN 1937769 B CN1937769 B CN 1937769B CN 200610113607 CN200610113607 CN 200610113607 CN 200610113607 A CN200610113607 A CN 200610113607A CN 1937769 B CN1937769 B CN 1937769B
- Authority
- CN
- China
- Prior art keywords
- frame
- intermediate frequency
- video
- coefficient
- watermark
- 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.)
- Expired - Fee Related
Links
Images
Landscapes
- Editing Of Facsimile Originals (AREA)
- Compression Or Coding Systems Of Tv Signals (AREA)
Abstract
The invention relates to an approach and equipment used for watermarking digital video signal. On the premise of ensuring the quality of video, it self-adoptively modulates partial intermediate frequency coefficient of each frame brightness component of video frequency to make the intermediate frequency energy of every frame brightness component of video frequency change obviously to inlay watermark information. Eigenvector is picked up through calculating the intermediate frequency magnitude relationship of adjacent frame's brightness component. It can judge whether watermark is added in the video content by using the relationship of the eigenvector and the watermark vector. On the premise of ensuring the quality of video, the invention can resist luminosity distortion, space in-phase distortion and the combination of these two kinds. It also can resist a certain degree of time in-phase distortion.
Description
Technical field
The present invention relates to the Digital Video Processing field, relate more specifically in digital video signal, add the method and apparatus of watermark, and detect the equipment whether digital video signal has embedded watermark.
Background technology
The fast development of the Internet makes the consumption demand of digital video constantly increase, and the various copyright problems of the thing followed also become increasingly conspicuous.Though in transmission course, can effectively protect digital content illegally not used, yet in case digital content arrives the end user and by after the legal deciphering, encryption technology just can not be carried out any protection to it by encryption technology.People press for a kind of new technology after digital content is decrypted, still can protect its content.Digital watermark technology then has the ability to satisfy these requirements.
As a kind of state-of-the-art technology of copyright protection, digital watermarking must solve a lot of problems before real the application.This one of them key problem is the robustness of watermark, experienced conventional signal processing exactly after, still can accurately detect the ability of watermark.For video, conventional signal processing comprises: several classes such as luminosity distortion (as diminishing compression, noise, filtering, contrast adjustment etc.), spatial synchronization distortion (as various geometric transformations such as cutting, scaling, translation, rotation, perspective, nonlinear transformation, spherical conversion, local geometric conversion), time synchronized distortion (as frame insertion, frame deletion, frame exchange, frame-rate conversion etc.) are handled.
In the achievement of watermark robustness research, the embedding of the spread-spectrum that Cox etc. propose and detect thought has solved the contradiction of fidelity and robustness aspect well, becomes the laying a foundation property thought of robustness watermark.In addition, generally improving the watermark robustness way of thinking commonly used also has: technology such as embedding, embed in the coefficient of known robustness in redundant embedding, the significant coefficient in perception.The general thoughts of antagonism time and spatial synchronization distortion then is exhaustive search, synchronization template, auto-correlation, constant watermark etc.But, these are the thought and the method for some basic raising watermark robustnesses, are not to solve the complete scheme that the robustness problem of a certain class distortion is resisted in watermark specially.
At present many robust watermarking methods are in order effectively to resist a certain class distortion in the three class distortions.As watermarked in the transform domain intermediate frequency coefficient, can resist slight luminosity distortion to a great extent; Utilize the constant embedding of RST subspace etc., opposing is because the slight spatial synchronization distortion that simple geometry conversion (RST) brings; As frame watermark keys maker, frame deletion, frame insertion, the frame that can resist at random exchange equal time synchronization attack etc. to Eugene T.L. with finite state machine.
In the present achievement in research about robust watermarking in the world, also there is part can resist the method for polytype problem of dtmf distortion DTMF simultaneously.As:
Jaap Haitsma has proposed the mean flow rate by each frame of modulation, the water mark method of embed watermark information frame by frame.This method can be resisted polytype spatial synchronization distortion preferably, but the ability of luminosity distortion such as anti-lossy compression method a little less than.Zhao Yao etc. improves Jaap Haitsma method, has improved the detection accuracy rate, but the ability of luminosity distortion such as anti-lossy compression method still a little less than.The water mark method based on energy that Song Maodong etc. propose can be resisted the spatial synchronization distortion that multiple geometric transformation causes, but can not resist the spatial synchronization distortion that cutting causes, and is general for the resistance of luminosity distortion.In addition, this method need be done whole discrete Fourier transform (DFT) to each two field picture, and amount of calculation is very big.
Niu Xia herd etc. propose along the watermarked method of time shaft, can resist the time synchronized distortion that the frame losing of certain frequency causes, can resist the spatial synchronization distortion that multiple geometric transformation causes, the luminosity distortion that causes for lossy compression method, resistance preferably being arranged also. but this method selects the method for watermark embedding point too complicated, and can not resist and all kinds ofly can cause watermark to embed point produces relativity shift in frame spatial synchronization distortions. Chen Zhenyong etc. have proposed to resist preferably simultaneously the water mark method of spatial synchronization distortion and slight time synchronized distortion. but change very fast when video pictures, when perhaps the video watermark sequence is longer, watermark correlation detection failure probability is very high. and so actual effective video watermark sequence is inevitable very short, this just causes the watermark space very little, suffer the attack of exhaustive mode easily, and the false alarm rate that detects is just than higher. in addition, this method for the ability of resisting the luminosity distortion a little less than.
The video watermark process based on " scene " that Hans Jung etc. propose can be resisted the spatial synchronization distortion that polytype time synchronized distortion and simple geometry conversion cause preferably.But this method amount of calculation is big, and video quality descends obviously.What Sun Jiande etc. proposed extracts the water mark method of video features based on ICA, can resist polytype time synchronized distortion preferably, can also resist common luminosity and attack, as compression etc.But, then there is not resistance for the spatial synchronization distortion.
In addition, based on the method for three-dimension varying, as based on methods such as 3D dct transform, 3D DWT conversion, the distortion that the processing of the part type in resisting this 3 big class distortion in varying degrees causes, but all have problems such as computational complexity height.Zhang Li and 3D Gabor conversion for frame deletion and video compression certain effect is arranged, but calculate too complicatedly, and almost can not resist the spatial synchronization distortion.
Above-mentioned argumentation is arranged as seen,, all can not resist spatial synchronization distortion and luminosity distortion effectively simultaneously though said method can be resisted a class or multiclass distortion to a certain extent.And when practical application, as video editing etc., video may experience multiple conversion and processing, and luminosity distortion, spatial synchronization distortion are simultaneous often.Therefore, press for research and can effectively resist the video watermark process of spatial synchronization distortion and luminosity distortion simultaneously.
Summary of the invention
At the defective that exists in the prior art, the invention discloses and a kind ofly new add watermark and detect method and the device whether digital video signal has embedded watermark for digital video signal.Can not resist the problem of spatial synchronization distortion and luminosity distortion effectively in order to solve the scheme that exists in the prior art.
A kind of water mark method that is used for digital video signal may further comprise the steps:
With the watermark information coded modulation is bipolarity binary watermarking vector W
r=(w (0) ..., w (n) ..., w (N-1)) and, N is the positive integer greater than 1;
At the start frame that selected watermark embeds, the inferior selecting video frame backward of N continuous is chosen the R frame at every turn, and R is the positive integer greater than 1;
Adjust the intermediate frequency energy of the luminance component of frame of video in R the frame of choosing for the n time, embed W
rN component, n=0～N-1.
Preferably, during the described intermediate frequency energy of adjusting the luminance component of frame of video in R the frame, the change direction of the intermediate frequency energy of the luminance component of each frame of video is incomplete same.
Preferably, the intermediate frequency energy of the luminance component of described adjustment frame of video may further comprise the steps:
(1) luminance component is split into equal-sized square image block;
(2) described image block is carried out discrete cosine transform DCT, obtain dct transform coefficient;
(3) in the dct transform coefficient of image block correspondence, select the part intermediate frequency coefficient;
(4) intermediate frequency coefficient of choosing is carried out the self adaptation adjustment, embed W
rN component;
(5), carry out reverse discrete cosine transform (IDCT) to adjusted each image block.
Preferably, in the described step (3), described intermediate frequency coefficient is the dct transform coefficient for an image block, when arranging by " it " font (ZigZag), is positioned at T AC coefficient of formation S to the, satisfies 0＜S＜T.
Preferably, in the described step (3), when selecting the DCT intermediate frequency coefficient, select according to the watermark embedded key, the coefficient of selection is no less than 1.
Preferably, in the described step (4), when the DCT intermediate frequency coefficient of choosing is carried out the self adaptation adjustment, adopt not embed or additivity embedding or the property taken advantage of embedded mode.
Preferably, in the described step (4), the DCT intermediate frequency coefficient of choosing is carried out self adaptation when adjusting, decides employing not embed or additivity embeds or the property taken advantage of embedded mode carries out according to following rule:
Remember that the current DCT intermediate frequency coefficient of choosing is that (i), wherein p is current frame number to DCT for p, m, and m is the sequence number of current block in this frame, and i is the sequence number of this coefficient in the ZigZag order, T
ABe positive constant, MultiT is greater than T
APositive number:
(i) absolute value is less than T for p, m as DCT
AThe time, take the mode that does not embed;
(i) absolute value is greater than or equal to T for p, m as DCT
A, during still less than MultiT, adopt the additivity embedded mode;
(p, m when i) absolute value is greater than or equal to MultiT, adopt the property taken advantage of embedded mode as DCT.
Preferably, MultiT is with the change in location of current video frame and image block, be designated as T (p, m).
Preferably, (p m) is two constant T to described T
mAnd T
MLinear combination, and T
A≤ T
m≤ T (p, m)≤T
M, T (p, m)=α (p, m) T
m+ (1-α (p, m)) T
M, (p m) is and the variable quantity of present frame with respect to former frame that current block is relevant with respect to the variable quantity of former frame co-located piece to its combination coefficient α.
Preferably, in the described step (4), when the DCT intermediate frequency coefficient of a certain frame luminance component correspondence chosen is adjusted,, change its adjusting range according to the intensity of variation of present frame luminance component with respect to the former frame luminance component.
Preferably, in the described step (4), when the DCT intermediate frequency coefficient of choosing certain image block is carried out the self adaptation adjustment,, change its adjusting range according to the amplitude of variation of current image block with respect to the image block of former frame co-located; Or
According to the model of the susceptibility of dct transform coefficient, change the adjusting range of this coefficient.
Preferably, in the described step (4), when the DCT intermediate frequency coefficient of choosing is carried out the self adaptation adjustment, guarantee that the DCT intermediate frequency energy of each image block in the frame changes toward same direction.
Preferably, to certain image block, the absolute value of all intermediate frequency coefficients that its DCT intermediate frequency energy is its dct transform and;
The intermediate frequency energy of the luminance component of one frame is the DCT intermediate frequency energy sum of all images piece of this frame luminance component.
Preferably, may further comprise the steps during watermark detection:
Select to have detected point frame, and from this frame, the inferior selecting video frame backward of N continuous is chosen the R frame at every turn;
Come n component of calculated characteristics vector, n=0～N-1 by the size of calculating the intermediate frequency energy of the luminance component of each frame in the R frame of choosing for the n time;
The correlation of calculated characteristics vector sum watermark vector;
Judge according to correlation size and detection threshold whether video content contains video watermark.
A kind of equipment that is used for watermark information is embedded into digital video signal comprises:
Being used for the watermark information coded modulation is bipolarity two-value vector W
rModulating device, W
r=(w (0) ..., w (n) ..., w (N-1)) and, N is the positive integer greater than 1;
Be used for the selecting video frame and carry out the frame selecting device that watermark embeds, at the start frame that selected watermark embeds, the inferior selecting video frame backward of N continuous is chosen the R frame at every turn, and R is the positive integer greater than 1;
Be used for watermarked watermark flush mounting, be used for adjusting the intermediate frequency energy of the luminance component of R the frame frame of video of choosing for the n time, embed n the component of Wr, n=0～N-1.
Preferably, described watermark flush mounting further comprises:
Be used for the frame of video luminance component is carried out piecemeal, carry out the brightness piecemeal converting means of discrete cosine transform then;
Be used for choosing the intermediate frequency coefficient choice device of DCT intermediate frequency coefficient according to watermark keys;
Be used for according to described frame of video with respect to the intensity of variation of former frame, current image block with respect to the intensity of variation of former frame co-located image block, the intermediate frequency energy and the current intermediate frequency coefficient order of magnitude of current image block, determine that current intermediate frequency coefficient adjusts the weighted volumes of weighted volumes and determine device;
Be used for that current intermediate frequency coefficient is adjusted weighted volumes and be added to adjusting device on the described video content;
Be used for and adjusted the piecemeal inverter that frame of video luminance component behind the intermediate frequency coefficient carries out reverse discrete cosine transform.
Preferably, the device of described definite intermediate frequency coefficient adjustment weighted volumes further comprises:
Be used to analyze video data, and calculate present frame with respect to former frame variable quantity and current block video data analytical equipment with respect to the variable quantity of former frame corresponding blocks;
Be used for according to present frame with respect to the former frame variable quantity, the overall situation of calculating present frame is adjusted the overall strength factor of intensity and is determined device;
Be used for according to the feature of current block and/or the current block variable quantity with respect to the former frame corresponding blocks, the part of calculating current image block is adjusted local strength's coefficient of intensity and is determined device;
Be used for choosing the sensitivity model of result and conversion coefficient, calculate the coefficient weighed intensities to be adjusted of coefficient weighed intensities to be adjusted and determine device according to intermediate frequency coefficient;
Be used for adjusting strength factor, local strength factor, the described coefficient weighed intensities to be adjusted adjusted according to the overall situation, finally the method for adjustment of definite intermediate frequency coefficient and the intermediate frequency coefficient method of adjustment and the weight coefficient of weight coefficient are determined device.
A kind of being used to detected the equipment whether digital video signal contains watermark, comprising:
Be used for the watermark information coded modulation is the modulating device of bipolarity binary watermarking vector;
Be used for from current detection point frame, N continuous time reads video successively, chooses the R frame at every turn, and the frame of video of choosing is carried out the frame of video selecting device of watermark detection, and wherein N is the positive integer greater than 1;
Be used for calculating the intermediate frequency energy calculation element of intermediate frequency energy of the luminance component of each frame of R frame of choosing for the n time, wherein n=0～N-1;
Be used for feature-vector extraction means according to the intermediate frequency Energy extraction detected characteristics vector of each frame that is calculated;
The correlation that is used for calculated characteristics vector sum watermark vector, and judge whether to contain the decision maker of watermark according to the correlation size.
The present invention is guaranteeing under the prerequisite that video quality does not have obviously to descend, the part intermediate frequency coefficient of the piecemeal dct transform of the luminance component by each frame of adaptive regulating video, make and the intermediate frequency energy generation marked change of luminance component of each frame of video come embed watermark information with this.The magnitude relationship of the intermediate frequency energy of the luminance component by calculating consecutive frame is extracted characteristic vector, utilizes the correlation of characteristic vector and watermark vector to judge whether video content has added watermark.
The present invention program not only can resist luminosity distortion, spatial synchronization distortion, and can resist the combination of this two classes distortion under the prerequisite of the original video quality of basic maintenance.This method also can be resisted time synchronized distortion to a certain degree in addition.
Description of drawings
Fig. 1 is used for watermark information is embedded into the structural representation of the equipment of digital video signal for the present invention;
Fig. 2 determines the structural representation of device for the weighted volumes of determining intermediate frequency coefficient among Fig. 1;
Whether Fig. 3 contains watermark for the present invention detects in the video device structure schematic diagram;
Fig. 4 is used for the schematic flow sheet of the water mark method of digital video signal for the present invention.
Embodiment
By as can be known to the research of digital watermark:
Come watermarkedly by adjusting the DCT intermediate frequency coefficient, can between the robustness of the fidelity of video quality and watermark, obtain balance preferably;
Watson has provided the model of two-dimensional dct transform coefficient sensitivity, for different coefficients, and the susceptibility difference.Utilize this characteristic, can self adaptation adjust the intensity of variation of each coefficient, and do not discovered by the people.
Adjacent video frames is the intelligible basis of video along the closely related property of time shaft.The frame of video that correlation is strong more, the correlation of the intermediate frequency energy that it is corresponding is also strong more, and the error of prediction is also more little mutually.Under the prerequisite that guarantees fidelity, can significantly revise the intermediate frequency energy of frame of video, thereby can utilize the symbol of the difference of adjacent video frames intermediate frequency energy to embed and detect watermark.
Because human eye storage characteristic, when the variation of adjacent pictures in the sequence of video images was violent, the resolving power of human eye can descend fast.It is human for the masking effect in the time domain of video pictures perception that Here it is.Therefore, guaranteeing under the non situation of human eye, changing violent picture than changing mild picture and can allow bigger variable quantity.Utilize this characteristic, can be according to the intensity of variation of each frame picture with respect to former frame, the change intensity of adaptively adjusting picture, and do not discovered by the people.
Below in conjunction with Figure of description the specific embodiment of the present invention is described.
As shown in Figure 1, the present invention program has comprised the equipment that watermark is embedded in the digital video signal:
Being used for the watermark information coded modulation is bipolarity two-value vector W
rModulating device 10, W wherein
r=(w (0) ..., w (n) ..., w (N-1)) and, N is the positive integer greater than 1;
Be used for the selecting video frame and carry out the frame of video selecting device 20 that watermark embeds, at the start frame that selected watermark embeds, the inferior selecting video frame backward of N continuous is chosen the R frame at every turn, and R is the positive integer greater than 1;
Be used for watermarked watermark flush mounting 30, be used for adjusting the intermediate frequency energy of the luminance component of R the frame frame of video of choosing for the n time, embed W
rN component, n=0～N-1.
Wherein, being used for watermarked watermark flush mounting 30 further comprises:
Be used for the frame of video luminance component is carried out piecemeal, carry out the brightness piecemeal converting means 310 of discrete cosine transform then;
Be used for choosing the intermediate frequency coefficient choice device 320 of DCT intermediate frequency coefficient according to watermark keys;
Be used for according to described frame of video with respect to the intensity of variation of former frame, current image block with respect to the intensity of variation of former frame co-located image block, the intermediate frequency energy and the current intermediate frequency coefficient order of magnitude of current image block, determine that current intermediate frequency coefficient adjusts the weighted volumes of weighted volumes and determine device 330;
Be used for being added to adjusting device 340 on the described video content with calculating adjustment amount;
Be used for and adjusted the piecemeal inverter 350 that frame of video luminance component behind the intermediate frequency coefficient carries out reverse discrete cosine transform.
The digital video signal that from digital video signal source, produces, enter frame of video selecting device 20, produce multi-frame video, in order to watermarked. each frame of selecting for video selecting device 20, handle through brightness piecemeal converting means 310, luminance component is carried out the piecemeal discrete cosine transform, then in conversion coefficient, requirement according to the intermediate frequency coefficient selecting device 320 that determines by watermark keys, select suitable intermediate frequency coefficient, adjust preparing. simultaneously according to the result that chooses of the output result of frame of video selecting device 20 and intermediate frequency coefficient selecting device 320, determine that by weighted volumes device 330 determines the weight of each intermediate frequency coefficient adjustment amount.
Watermark information to be embedded obtains watermark vector after over-modulating device 10 modulation.Adjusting device 340 is determined the conversion coefficient that weight that device 330 obtains and brightness piecemeal converting means 310 obtain according to this vector, weighted volumes, determines the adjustment amount of each intermediate frequency coefficient, and intermediate frequency coefficient is adjusted.Adjusted result carries out inverse transformation through piecemeal inverter 350, obtains the digital video signal of embed watermark information, is saved on the suitable storage medium 40.
As shown in Figure 2, the weighted volumes of determining intermediate frequency coefficient among Fig. 1 determines that device 330 further comprises:
Be used to analyze video data, and calculate present frame with respect to former frame variable quantity and current block video data analytical equipment 331 with respect to the variable quantity of former frame corresponding blocks;
Be used for according to present frame with respect to the former frame variable quantity, the overall situation of calculating present frame is adjusted the overall strength factor of intensity and is determined device 332;
Be used for according to the feature of current block and/or the current block variable quantity with respect to the former frame corresponding blocks, the part of calculating current image block is adjusted local strength's coefficient of intensity and is determined device 333;
Be used for choosing the sensitivity model of result and conversion coefficient, calculate the coefficient weighed intensities to be adjusted of coefficient weighed intensities to be adjusted and determine device 334 according to intermediate frequency coefficient;
Be used for determining that according to overall strength factor the overall situation that device 332 is exported is adjusted strength factor, local strength's coefficient determines that strength factor is adjusted in the part of device 333 outputs and coefficient weighed intensities to be adjusted is determined the single adjustment intensity weighted coefficient that device 334 is exported, finally the method for adjustment of definite intermediate frequency coefficient and the intermediate frequency coefficient method of adjustment and the weight coefficient of weight coefficient are determined device 335.
As shown in Figure 3, be the equipment that whether contains watermark in the detection video of correspondence of the present invention, comprising:
Be used for the watermark information coded modulation is the modulating device 10 of bipolarity binary watermarking vector;
Be used for the frame of video selecting device 50 that the selecting video frame carries out watermark detection;
Be used to calculate the intermediate frequency energy calculation element 60 of the intermediate frequency energy of frame of video luminance component;
Be used for feature-vector extraction means 70 according to the intermediate frequency Energy extraction detected characteristics vector of each frame that is calculated;
Be used for correlation, and judge whether to contain the decision maker 80 of watermark according to the correlation size calculated characteristics vector sum watermark vector.
According to the said equipment, the digital video signal that produces in the digital video signal source to be detected through choosing of frame of video selecting device 50, obtains video segment to be detected.Calculate the intermediate frequency energy of each frame of video in this video segment through intermediate frequency energy calculation element 60.Feature-vector extraction means 70 extracts the characteristic vector of current video fragment according to the intermediate frequency energy of each frame of video.Simultaneously watermark information is modulated via modulating device 10, obtained watermark vector.Decision maker 80 judges according to characteristic vector and watermark vector whether the current video fragment contains watermark.
Illustrate that below in conjunction with said apparatus the present invention is used for the water mark method of digital video signal.As shown in Figure 4, be the schematic flow sheet of this method, as seen from the figure, mainly may further comprise the steps:
Step 401, be bipolarity binary watermarking vector with the watermark information coded modulation;
This step is modulated to bipolarity binary watermarking vector W with watermark information to be embedded through over-modulating device 10
r=(w (0) ..., w (n) ..., w (N-1)), wherein w (n) ∈ 1,1}, n=0～N-1; N is the positive integer greater than 1.
Step 402, the start frame that embeds in selected watermark, N continuous be the selecting video frame time backward, chooses the R frame at every turn, and R is the positive integer greater than 1;
In the digital video signal of digital video signal source generation, select suitable frame by frame of video selecting device 20, to be used for the embedding of watermark.During selection, the at first selected initial point frame that embeds.From embedding point frame, N continuous is inferior to read frame of video successively backward, in the embodiment of the invention, reads the R=3 frame at every turn then.For the 3 frame videos that obtain for the n time, be prepared for n the component w (n) of watermarked vector, wherein n=0～N-1.
The intermediate frequency energy of the luminance component of frame of video embeds W in R the frame that step 403, adjustment are chosen for the n time
rN component, n=0～N-1.
For the frame of video that frame of video selecting device 20 is chosen, carry out conversion through brightness piecemeal converting means 310.During conversion, the luminance component of selecting frame of video can be carried out the dct transform of 8 * 8 piecemeals, the size of image block is 8 * 8.Unless following in question dct transform specializes, all be meant the dct transform by 8 * 8 piecemeals of the luminance component of video, but the present invention program does not get rid of the image block method of using other, uses other mapping modes equally can realize the present invention program yet.
Obtain DCT after the conversion and divide block conversion coefficient, be used for adjusting through intermediate frequency coefficient selecting device 320 selected part intermediate frequency coefficients.
According to watermark keys, the dct transform coefficient of selecting 8 * 8 is according to i coefficient of ZigZag order, i=S wherein ..., T, 0＜S＜T＜32.
In the such scheme, determine that the final weighted volumes of each coefficient adopts following method to carry out:
For the 3 frame videos that obtain for the n time, be prepared for n the component w (n) of watermarked vector, n=0～N-1 wherein, establishing its intermediate frame is p with respect to the frame number of current embedding starting point, and p=n * 3+1 is arranged.
Set up departments several DCT (p, m, i) Dui Ying weight coefficient be Ω (i), its computational methods are as follows for p, m:
Ω(p，m，i)＝λ(i)·φ(p，m，i)
λ(i)＝max(1，DCT_Freq_Sensitivity(i)×Scale
D)
Wherein DCT_Freq_Sensitivity (i) is i the sensitivity value that DCT frequency sensitive kilsyth basalt is pressed the ZigZag order, Scale
DBe an adjustment factor, max () is for getting peaked function.T
ABe a positive constant, (p m) is two constant T to T
mAnd T
MLinear combination, and T
A≤ T
m≤ T (p, m)≤T
MT (p, m) computational methods are as follows:
T(p，m)＝α(p，m)·T
m+(1-α(p，m))·T
M
Wherein (p m) is combination coefficient to α, and (p, m) (p-1, (p, intermediate frequency energy m) is determined for variable quantity m), piece B with respect to B with respect to variable quantity, the piece B of p-1 by frame p for it.
Determine the weighted volumes that device 330 calculates according to weighted volumes, the conversion coefficient that brightness piecemeal converting means 310 obtains, the watermark vector that modulating device 340 obtains, adjusting device 340 will calculate the actual adjustment amount of each coefficient, and adjust, and method of adjustment is as follows:
If DCT (p, m, i) be the dct transform coefficient of m 8 * 8 of p frame according to i coefficient of zigzag order, coefficient DCT (p, m, i) adjusted new value be DCT (p, m, i) ', its computational methods are:
DCT(p，m，i)′＝DCT(p，m，i)+Δ(p，m，i)
Wherein Δ (p, m, i) amount for a change, Δ (p, m, i)=Ω (p, m, i) * w (n).
In order to keep the not sentience of watermark, keep watermark to embed the quality of rear video, we with interval [C (and p, i), C (p, i)] come that (p, m i) carry out restricted intercepting to Δ.Wherein C (p, i) can calculate as follows:
Wherein D (p) is the variable quantity of p frame with respect to the p-1 frame.
For the adjustment of the intermediate frequency coefficient of the dct transform of video p-1 frame, can adopt following method:
Replace p in the above-mentioned steps with p-1, and replace the w (n) in the above-mentioned steps with-w (n), according to above-mentioned steps, the intermediate frequency coefficient of the dct transform of adjustment video p-1 frame.
For the adjustment of the intermediate frequency coefficient of the dct transform of video p+1 frame, can adopt following method:
Replace p in the above-mentioned steps with p+1, and replace the w (n) in the above-mentioned steps with-w (n), according to above-mentioned steps, the intermediate frequency coefficient of the dct transform of adjustment video p+1 frame.
Adjusting device 340 adjusted results, carry out video after conversion obtains adding watermark through piecemeal inverter 350, transform method is the adjusted DCT coefficient to luminance component, carries out 8 * 8 piecemeal inverse discrete cosine transforms.
The present invention program, can comprise the steps: for the digital video signal that whether contains watermark to be detected at receiving terminal
Select suitable frame of video by frame of video selecting device 50;
From current detection point frame, N continuous is inferior backward successively reads video, in the embodiment of the invention, adopts and reads 3 frames at every turn.For the 3 continuous frame videos that read for the n time, select intermediate frame, remember that its frame number with respect to the current detection starting point is p, p=n * 3+1 is then arranged.N=0～N-1 wherein.
For the frame of video of choosing, calculate the intermediate frequency energy through intermediate frequency energy calculation element 60, adopt following method when calculating the intermediate frequency energy:
If M is 8 * 8 a number in video one frame, B (p m) is m piece of p frame, DCT (p, m, i) be the dct transform coefficient of individual 8 * 8 of m of p frame according to i coefficient of zigzag order, i=S wherein ..., T, 0＜S＜T＜32.If E
m(p) be the intermediate frequency energy of video p frame, computational methods are as follows:
Calculated after the intermediate frequency energy, extract characteristic vector with feature-vector extraction means 70.The method of extracting characteristic vector is as follows:
Characteristic vector V=(v (0), v (1) ... v (n) ..., v (N-1)) the computational methods of component v (n) as follows:
Wherein, ε
0Be a positive constant, ε (n) computational methods are as follows:
According to the characteristic vector of feature-vector extraction means 70 output, and by the watermark vector of modulating device 10 outputs, decision maker 80 judges whether the video-frequency band of choosing contains watermark, and judgment mode is as follows:
At first calculated characteristics vector V and watermark vector W
rNormalization correlation τ.
If τ
NcBe detection threshold.If τ 〉=τ
Nc, show that then detected this section video has watermark; Otherwise detected this section video does not have watermark.
The results showed, scheme disclosed in this invention is dwindled geometric ratio amplification with 4 * 4 etc. for non-geometric ratio scaling, perspective, 1/4 * 1/4 geometric ratio and has all been reached 100% detection accuracy rate, geometric transformation for combination as cutting and non-geometric ratio scaling, reaches 100% detection accuracy rate equally.This is that additive method is unapproachable, and this explanation the present invention program can effectively resist the spatial synchronization distortion.
Scheme disclosed in this invention is for the luminosity distortion of 4 kinds of main types: the detection accuracy rate of noise, changes in amplitude, filtering, lossy compression method has all reached 100%.Especially lossy compression method, under the compression of VCD level quality (video that is lower than the VCD quality generally is that people are unacceptable), detecting accuracy rate still is 100%, this is that the additive method method is difficult to realize.
When spatial synchronization distortion and luminosity distortion took place simultaneously, scheme disclosed in this invention still can reach 100% detection accuracy rate, and this also is that additive method is difficult to accomplish.
In addition, video watermark process disclosed in this invention also has certain robustness for the time synchronized distortion that random frame is inserted and frame deletion causes.
Obviously, those skilled in the art can carry out various changes and modification to the present invention and not break away from the spirit and scope of the present invention.Like this, if of the present invention these are revised and modification belongs within the scope of claim of the present invention and equivalent technologies thereof, then the present invention also is intended to comprise these changes and modification interior.
Claims (18)
1. water mark method that is used for digital video signal may further comprise the steps:
With the watermark information coded modulation is bipolarity binary watermarking vector W
r=(w (0) ..., w (n) ..., w (N-1)) and, wherein, { 1,1}, N are the positive integer greater than 1 to w (n) ∈;
At the start frame that selected watermark embeds, the inferior selecting video frame backward of N continuous is chosen the R frame at every turn, and R is the positive integer greater than 1;
Adjust the intermediate frequency energy of the luminance component of frame of video in R the frame of choosing for the n time, embed W
rN component, n=0～N-1.
2. the method for claim 1 is characterized in that:
During the described intermediate frequency energy of adjusting the luminance component of frame of video in R the frame, the change direction of the intermediate frequency energy of the luminance component of each frame of video is incomplete same.
3. the method for claim 1 is characterized in that, the intermediate frequency energy of the luminance component of described adjustment frame of video may further comprise the steps:
(1) luminance component is split into equal-sized square image block;
(2) described image block is carried out discrete cosine transform DCT, obtain dct transform coefficient;
(3) in the dct transform coefficient of image block correspondence, select the part intermediate frequency coefficient;
(4) intermediate frequency coefficient of choosing is carried out the self adaptation adjustment, embed W
rN component;
(5), carry out reverse discrete cosine transform (IDCT) to adjusted each image block.
4. method as claimed in claim 3 is characterized in that, in the described step (3), described intermediate frequency coefficient is the dct transform coefficient for an image block, when arranging, be positioned at T AC coefficient of formation S to the, satisfy 0＜S＜T by " it " font (ZigZag).
5. method as claimed in claim 3 is characterized in that, in the described step (3), when selecting the DCT intermediate frequency coefficient, selects according to the watermark embedded key, and the coefficient of selection is no less than 1.
6. method as claimed in claim 3 is characterized in that, in the described step (4), when the DCT intermediate frequency coefficient of choosing is carried out the self adaptation adjustment, adopts additivity to embed or the property taken advantage of embedded mode.
7. method as claimed in claim 6 is characterized in that, in the described step (4), when the DCT intermediate frequency coefficient of choosing is carried out the self adaptation adjustment, decides embedding of employing additivity or the property taken advantage of embedded mode to carry out according to following rule:
Remember that the current DCT intermediate frequency coefficient of choosing is that (i), wherein p is current frame number to DCT for p, m, and m is the sequence number of current block in this frame, and i is the sequence number of this coefficient in the ZigZag order, T
ABe positive constant, MultiT is greater than T
APositive number:
(i) absolute value is greater than or equal to T for p, m as DCT
A, during still less than MultiT, adopt the additivity embedded mode;
(P, m when i) absolute value is greater than or equal to MultiT, adopt the property taken advantage of embedded mode as DCT.
8. method as claimed in claim 7 is characterized in that, MultiT is with the change in location of current video frame and image block, be designated as T (p, m).
9. method as claimed in claim 8 is characterized in that, (p m) is two constant T to described T
mAnd T
MLinear combination, and T
A≤ T
m≤ T (p, m)≤T
M, T (p, m)=α (p, m) T
m+ (1-α (p, m)) T
M, (p m) is and the variable quantity of present frame with respect to former frame that current block is relevant with respect to the variable quantity of former frame co-located piece to its combination coefficient α.
10. method as claimed in claim 3, it is characterized in that, in the described step (4), when the DCT intermediate frequency coefficient of a certain frame luminance component correspondence chosen is adjusted, according to the intensity of variation of present frame luminance component, change its adjusting range with respect to the former frame luminance component.
11. method as claimed in claim 3, it is characterized in that, in the described step (4), when the DCT intermediate frequency coefficient of choosing certain image block is carried out the self adaptation adjustment, according to the amplitude of variation of current image block, change its adjusting range with respect to the image block of former frame co-located; Or
According to the model of the susceptibility of dct transform coefficient, change the adjusting range of this coefficient.
12. method as claimed in claim 3 is characterized in that, in the described step (4), when the DCT intermediate frequency coefficient of choosing is carried out the self adaptation adjustment, guarantees that the DCT intermediate frequency energy of each image block in the frame changes toward same direction.
13. as any described method of claim 3 to 12, it is characterized in that,
To certain image block, the absolute value of all intermediate frequency coefficients that its DCT intermediate frequency energy is its dct transform and;
The intermediate frequency energy of the luminance component of one frame is the DCT intermediate frequency energy sum of all images piece of this frame luminance component.
14. the method for claim 1 is characterized in that, may further comprise the steps during watermark detection:
Select to have detected point frame, and from this frame, the inferior selecting video frame backward of N continuous is chosen the R frame at every turn;
Come n component of calculated characteristics vector, n=0～N-1 by the size of calculating the intermediate frequency energy of the luminance component of each frame in the R frame of choosing for the n time;
The correlation of calculated characteristics vector sum watermark vector;
Judge according to correlation size and detection threshold whether video content contains video watermark.
15. an equipment that is used for watermark information is embedded into digital video signal comprises:
Being used for the watermark information coded modulation is bipolarity two-value vector W
rModulating device, W
r=(w (0) ..., w (n) ..., w (N-1)) and, wherein, { 1,1}, N are the positive integer greater than 1 to w (n) ∈;
Be used for the selecting video frame and carry out the frame selecting device that watermark embeds, at the start frame that selected watermark embeds, the inferior selecting video frame backward of N continuous is chosen the R frame at every turn, and R is the positive integer greater than 1;
Be used for watermarked watermark flush mounting, be used for adjusting the intermediate frequency energy of the luminance component of R the frame frame of video of choosing for the n time, embed n the component of Wr, n=0～N-1.
16. equipment as claimed in claim 15 is characterized in that, described watermark flush mounting further comprises:
Be used for the frame of video luminance component is carried out piecemeal, carry out the brightness piecemeal converting means of discrete cosine transform then;
Be used for choosing the intermediate frequency coefficient choice device of DCT intermediate frequency coefficient according to watermark keys;
Be used for according to described frame of video with respect to the intensity of variation of former frame, current image block with respect to the intensity of variation of former frame co-located image block, the intermediate frequency energy and the current intermediate frequency coefficient order of magnitude of current image block, determine that current intermediate frequency coefficient adjusts the weighted volumes of weighted volumes and determine device;
Be used for that current intermediate frequency coefficient is adjusted weighted volumes and be added to adjusting device on the video content;
Be used for and adjusted the piecemeal inverter that frame of video luminance component behind the intermediate frequency coefficient carries out reverse discrete cosine transform.
17. equipment as claimed in claim 16 is characterized in that, the device that described definite intermediate frequency coefficient is adjusted weighted volumes further comprises:
Be used to analyze video data, and calculate present frame with respect to former frame variable quantity and current block video data analytical equipment with respect to the variable quantity of former frame corresponding blocks;
Be used for according to present frame with respect to the former frame variable quantity, the overall situation of calculating present frame is adjusted the overall strength factor of intensity and is determined device;
Be used for according to the feature of current block and/or the current block variable quantity with respect to the former frame corresponding blocks, the part of calculating current image block is adjusted local strength's coefficient of intensity and is determined device;
Be used for choosing the sensitivity model of result and conversion coefficient, calculate the coefficient weighed intensities to be adjusted of coefficient weighed intensities to be adjusted and determine device according to intermediate frequency coefficient;
Be used for adjusting strength factor, local strength factor, the described coefficient weighed intensities to be adjusted adjusted according to the overall situation, finally the method for adjustment of definite intermediate frequency coefficient and the intermediate frequency coefficient method of adjustment and the weight coefficient of weight coefficient are determined device.
18. one kind is used to detect the equipment whether digital video signal contains watermark, comprises:
Be used for the watermark information coded modulation is the modulating device of bipolarity binary watermarking vector;
Be used for from current detection point frame, N continuous time reads video successively, chooses the R frame at every turn, and the frame of video of choosing is carried out the frame of video selecting device of watermark detection, and wherein N is the positive integer greater than 1, and R is the positive integer greater than 1;
Be used for calculating the intermediate frequency energy calculation element of intermediate frequency energy of the luminance component of each frame of R frame of choosing for the n time, wherein n=0～N-1;
Be used for feature-vector extraction means according to the intermediate frequency Energy extraction detected characteristics vector of each frame that is calculated;
The correlation that is used for calculated characteristics vector sum watermark vector, and judge whether to contain the decision maker of watermark according to the correlation size.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 200610113607 CN1937769B (en) | 2006-10-09 | 2006-10-09 | Water print method and device for digital video signal |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 200610113607 CN1937769B (en) | 2006-10-09 | 2006-10-09 | Water print method and device for digital video signal |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN1937769A CN1937769A (en) | 2007-03-28 |
| CN1937769B true CN1937769B (en) | 2010-05-12 |
Family
ID=37954987
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN 200610113607 Expired - Fee Related CN1937769B (en) | 2006-10-09 | 2006-10-09 | Water print method and device for digital video signal |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN1937769B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102238421A (en) * | 2010-04-29 | 2011-11-09 | 华为终端有限公司 | Multi-path video synchronous transmission method, device and system |
| CN106454510B (en) * | 2016-10-31 | 2019-07-16 | 广东工业大学 | A kind of original video watermark handling method and device |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6792129B1 (en) * | 1999-08-25 | 2004-09-14 | Sharp Laboratories Of America | Digital watermarking in a perceptually uniform domain |
| CN1600033A (en) * | 2001-12-05 | 2005-03-23 | 皇家飞利浦电子股份有限公司 | Embedding a watermark in an image signal |
| CN1608376A (en) * | 2001-06-21 | 2005-04-20 | 皇家菲利浦电子有限公司 | Embedding and detection of watermark in a motion image signal |
| EP1684265A1 (en) * | 2005-01-21 | 2006-07-26 | Unlimited Media GmbH | Method of embedding a digital watermark in a useful signal |
-
2006
- 2006-10-09 CN CN 200610113607 patent/CN1937769B/en not_active Expired - Fee Related
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6792129B1 (en) * | 1999-08-25 | 2004-09-14 | Sharp Laboratories Of America | Digital watermarking in a perceptually uniform domain |
| CN1608376A (en) * | 2001-06-21 | 2005-04-20 | 皇家菲利浦电子有限公司 | Embedding and detection of watermark in a motion image signal |
| CN1600033A (en) * | 2001-12-05 | 2005-03-23 | 皇家飞利浦电子股份有限公司 | Embedding a watermark in an image signal |
| EP1684265A1 (en) * | 2005-01-21 | 2006-07-26 | Unlimited Media GmbH | Method of embedding a digital watermark in a useful signal |
Non-Patent Citations (1)
| Title |
|---|
| JP特开2004-172758A 2004.06.17 |
Also Published As
| Publication number | Publication date |
|---|---|
| CN1937769A (en) | 2007-03-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Pereira et al. | Fast robust template matching for affine resistant image watermarks | |
| Dittmann et al. | Content-based digital signature for motion pictures authentication and content-fragile watermarking | |
| AU758814B2 (en) | A method for inserting and detecting electronic watermark data into a digital image and a device for the same | |
| CN102880999B (en) | Color image digital watermark method capable of resisting stamping, printing and scanning processes | |
| CN108682425B (en) | Robust digital audio watermark embedding system based on constant watermark | |
| CN104268823A (en) | Digital watermark algorithm based on image content | |
| CN100481867C (en) | A digital watermark embedding and extraction algorithm for space domain and small wave domain double sub-set | |
| CN104053074B (en) | Video watermarking method based on depth image and Otsu segmentation | |
| CN110232651A (en) | A kind of reversible information hidden method based on block prediction and difference histogram shifting | |
| CN104766269A (en) | Spread transform dither modulation watermarking method based on JND brightness model | |
| CN107067363A (en) | A kind of non-blind robust Spatial Domain Image Watermark method constant based on pixel order relation | |
| CN102117474B (en) | Digital picture watermark embedding and detecting method and device | |
| Khan et al. | On hiding secret information in medium frequency DCT components using least significant bits steganography | |
| CN107346528A (en) | A kind of distorted image detection method based on the reversible fragile watermark of digraph | |
| CN105741222B (en) | A kind of steganography information locating method based on the estimation of pixel subset insertion rate | |
| CN1937769B (en) | Water print method and device for digital video signal | |
| CN104123693A (en) | Multi-functional digital watermarking method for three-dimensional picture | |
| CN103325084B (en) | A kind of stereo-picture vulnerable watermark method just can perceiveing distortion based on associating | |
| Shi et al. | Tailoring reversible data hiding for 3D synthetic images | |
| Tao et al. | Robust digital image watermarking in curvelet domain | |
| CN101853483B (en) | Printing and scanning resilient watermarking method based on mathematical modeling | |
| CN100512440C (en) | Water mark method and device for digital video signal and detecting method and device | |
| Behravan et al. | Introducing a new method of image reconstruction against crop attack using sudoku watermarking algorithm | |
| Leng et al. | High payload data hiding based on just noticeable distortion profile and LSB substitution | |
| CN102497602A (en) | Digital watermarking processing method and device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20100512 Termination date: 20191009 |