CN1527257A - Object-oriented self-synchronizing water mark detecting method for multimedia product - Google Patents
Object-oriented self-synchronizing water mark detecting method for multimedia product Download PDFInfo
- Publication number
- CN1527257A CN1527257A CNA031510388A CN03151038A CN1527257A CN 1527257 A CN1527257 A CN 1527257A CN A031510388 A CNA031510388 A CN A031510388A CN 03151038 A CN03151038 A CN 03151038A CN 1527257 A CN1527257 A CN 1527257A
- Authority
- CN
- China
- Prior art keywords
- watermark
- information
- ellipse
- inertia
- water mark
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Abstract
The object-oriented self-synchronizing water mark detecting method for multimedia product adopts main shaft conversion to constitute inertial ellipse to describe azimuth and size information of object in any shape and analyzes and predict the geometric attack the target in the transmission channel suffers before embedding the water mark and after suffering attack to realize the fast self-synchronized calibration of the object and obtain the self-synchronization of water mark detection. By means of combining shape-adaptive cosine transform a human visual shading model, disturbed water mark information is embedded into target of different shapes to highest strength. Wiener filtering is used to predict the initial number of the target and this makes it possible to need no participation of initial image in water mark detection. The method of the present invention has improved robustness against geometric attack and may be used in the copy right protection of video object.
Description
Technical field:
The present invention relates to OO motor synchronizing method of detecting watermarks in a kind of media product; utilize the motor synchronizing watermark detection of technology realization object under geometric attack such as the ellipse of inertia; realize the copyright protection of arbitrary shape object in the media product, belong to field of information security technology.
Background technology:
Along with network technology and development of multimedia technology, the media product of digital forms such as image, text, Voice ﹠ Video has obtained propagating widely and using, and multimedia communication becomes the important means of information interchange between the people gradually.But; digital product is easy to be replicated, handles and discloses propagation; the bootlegger utilizes legitimate right that these characteristics of digital product destroy manufacturer and user to obtain individual interest just, has caused the Copyright Protection of digital information safety problem and media product thus.Information security not only relates to national interests with secret, and with commercial activity and individual great relation is arranged all, and therefore, the multi-media information security technology day by day becomes the focus that people are concerned about, also is the research focus in the current infotech.
Digital watermark technology is exactly secret digital signal or the pattern with identifiability that embed in media products such as image, sound, document, video, with cradle, the owner, content, the right to use, sequence number, right, integrality or the terminal point etc. that are used for discerning or identifying digital product.Because digital watermark is the secret data such as explanation picture material, parameter and note that embed under the situation that does not influence the image perceptual quality; not only can solve the Copyright Protection of media product, also can be multimedia data file authentication, false proof, anti-tamper, ensure that data security and aspect such as complete provide effective technical means.
The digital watermark technology that is used for copyright protection requires that various signal processings are had very strong repellence, even under the situation of data distortion, still can keep the integrality of watermark and the accuracy of discriminating.Most of water mark methods are for general image processing operations, for example compression, filtering etc. have stronger robustness, still, (as convergent-divergent, rotation, displacement etc.) but seem powerless for geometric attack, and often very little geometric distortion just can be routed the watermark detection capability of method.The field that the method for detecting watermarks general using of existing opposing geometric attack remains unchanged to geometric transformation (, paying Li Ye-Malin's conversion (Fourier-Mellin), invariant moments etc.) as discrete fourier transition (DFT); Or utilize the autocorrelation function of watermark itself; Or use the added pattern that affined transformation is had a synchronization detection capability to wait and realize motor synchronizing.But these methods mostly propose at rest image, and watermark information is evenly dispersed in the entire image, the not special local feature of considering image.
New video encoding standard MPEG-4 provides a kind of OO image expression mode, can improve video compression ratio greatly, satisfies the requirement of more wide multimedia application.The formulation of new standard, also new problem has been proposed for the copyright protection of media product: copyright how to protect the arbitrary shape object? at this problem, the digital watermark method that Swanson at first proposed based on discrete cosine transform (DCT) in 1997, but this method belongs to the frame video watermark process in fact, do not consider the arbitrary shape object encoded question (" Object-based transparent video watermarking; " IEEE workshop on Multimediasignal processing, Princeton, New Jersey, USA, pp:369-374,1997); Kim has compared that OO gray scale is filled digital watermark and based on the digital watermark of form adaptive cosine transform (SA-DCT), think robustness based on the digital watermark of form adaptive cosine transform be better than gray scale fill digital watermark (" Anobject-based video watermarking; " International Conference on ConsumerElectronics, pp:100-102,1999); Wu proposed based on the water mark method of form adaptive wavelet transformation (" Object-based multiresolution watermarking of images and video; " IEEEInternational Symposium on Circuits and Systems, Vol.1, pp:212-215,2000); But these methods are not all considered the motor synchronizing of watermark behind the geometric attack and are detected problem.In MPEG-4, because object video can but can not cause tangible vision distortion by hack, so that this problem becomes is particularly important.
Summary of the invention:
The objective of the invention is at the deficiencies in the prior art and actual needs, OO motor synchronizing method of detecting watermarks in a kind of new media product is provided, even suffered geometric attacks such as convergent-divergent, rotation, displacement at object after, still can detect the watermark information of embedded object.
For realizing such purpose, the present invention proposes to utilize principal axis transformation to find the solution the eigenwert and the proper vector of arbitrary shape object covariance, side information---the ellipse of inertia of structure object, the geometric attack that comes forecasting object in transmission channel, to experience, realize the quick motor synchronizing calibration of object, thereby obtain the motor synchronizing of watermark detection.
At the watermark built-in end, at first utilize the ellipse of inertia that the initial object of embed watermark is not carried out content analysis, to obtain the side information of object; Utilize the image scrambling technology with significant watermark information scramble then, utilize the form adaptive cosine transform at last, watermark information behind the scramble is embedded in the object of arbitrary shape, realize that watermark embeds in conjunction with the human vision shadow model.
At the watermark detection end, at first the object of being attacked is carried out ellipse of inertia analysis, obtain the synchronizing information of object, and compare, to carry out the motor synchronizing calibration of object with object edges information; Utilize Wiener filtering to operate then and predict the initial object data; According to the inverse process of watermark embed process, from object, extract watermark at last, and, carry out correlativity with original watermark information and judge through after the scramble recovery.
The concrete steps of the inventive method are as follows:
1, object edges information obtains
The present invention utilizes the principal axis transformation in the pattern-recognition, derives to have constructed the ellipse of inertia that is used for describing arbitrary shape object orientation and size information, realizes the quick motor synchronizing calibration of object behind the geometric attack.Therefore, before digital watermarking embeds, at first the object of embed watermark is not carried out ellipse of inertia analysis, eigenwert and proper vector according to the object covariance matrix, obtain side informations such as the direction of center of gravity, main shaft of object and size, and be delivered to the watermark detection end, in order to realize the object motor synchronizing calibration in the watermark detection process.
2, the embedding of digital watermarking
Watermark information is a significant bianry image, consider the tolerance of human eye, distribute for making watermark signal more approach random noise for random noise, and the bit that is damaged when disperseing watermark detection, adopt the Arnold conversion with the watermark scramble, carry out watermark again and embed.
It is the linear adder strategy that watermark embeds rule, and watermark signal is hidden in the special frequency channel of image according to the shape and the vision shadow model of object.The object of arbitrary shape is divided by 8 * 8 sub-pieces, and, image block is divided into background piece and object piece according to the shape information of object.The object piece can be divided into object home block and object bounds piece again, and the object home block is encoded with normal scatter cosine transform DCT, and the object bounds piece realizes that with form adaptive discrete cosine transform SA-DCT coding watermark embeds.Because human eye is responsive to the noise ratio of low-frequency range, and high band information might be lost in compression process, so watermark information should embed the Mid Frequency of DCT coefficient, to obtain the balance of robustness and invisibility.The embedment strength of watermark can be under sightless prerequisite by human vision shadow model adaptive determining, the embed watermark of maximum intensity.At last,, carry out reverse discrete cosine transform, obtain the object of embed watermark according to the shape information of object.
3, the motor synchronizing of object calibration
In the testing process of digital watermarking, the motor synchronizing information of at first carrying out object detects.Utilize the ellipse of inertia technical Analysis information such as center of gravity, major axes orientation and length of object under fire, and the object edges information before embedding with watermark compares, to adjust the synchronism that object obtains watermark detection.It is pointed out that the center of gravity of destination object should be complementary with the center of its ellipse of inertia in the process of object calibration, not only, also divide for the image block of controlling object in the watermark extraction process in order to carry out accurate object calibration.
4, the detection of digital watermarking
The testing process of digital watermarking is the inverse process of digital watermark embed process basically.For realizing that watermark detection does not need original image, utilize Wiener filtering to come the raw value of forecasting object.Embed the zone in watermark, the difference symbol of the DCT value of destination object and the DCT value of forecasting object has determined the watermark predicted value at this place.After the watermark that dopes recovered through the Arnold scramble, the similarity of calculating and original watermark signal was with the quality of the watermark information that objective evaluation was obtained.Correlativity by prediction watermark and original watermark judges relatively whether watermark exists, if similarity greater than the preset detection threshold value, is that watermark exists with regard to decidable.
The present invention has significant beneficial effect.The demand angle of the present invention's object video copyright protection from new video standard MPEG-4; motor synchronizing method of detecting watermarks towards the arbitrary shape object has been proposed; after can having suffered geometric attacks such as convergent-divergent, rotation, displacement at object, still realize its copyright protection.
At the watermark built-in end, at first utilize the ellipse of inertia that the initial object of embed watermark is not carried out content analysis, obtaining the side information of object, and be delivered to watermark detection and bring in the motor synchronizing watermark detection that realizes object behind the geometric attack; Utilize the image scrambling technology with significant watermark information scramble then, so that watermark signal more distributes near random noise, and the bit that is damaged when disperseing watermark detection, thereby improve the robustness of watermarking algorithm; The form adaptive cosine transform is embedded into watermark information behind the scramble in the object of arbitrary shape in conjunction with the human vision shadow model, and under sightless prerequisite, the maximum intensity embed watermark is to improve the robustness that the watermarking algorithm opposing is attacked.
At the watermark detection end, at first, obtain the synchronizing information of object, and after comparing with object edges information, realize the motor synchronizing calibration of object being carried out ellipse of inertia analysis by the object attacked; Utilize Wiener filtering to operate then and predict the initial object data, so watermark detection do not need the participation of original image, thereby enlarged the usable range of watermarking algorithm; And, from object, extract watermark according to the inverse process of watermark embed process, and after recovering through scramble, carry out correlativity with original watermark information and judge, judge with objective whether watermark exists.
The present invention has improved the robustness of OO digital watermark method to geometric attack; be mainly used in the copyright protection of object video among the MPEG-4; also can be applicable in the rest image water mark method design, also object-based image retrieval is had potential application foreground based on feature.
Description of drawings:
Fig. 1 is the FB(flow block) of object-oriented motor synchronizing method of detecting watermarks of the present invention.
As shown in Figure 1,, at first utilize principal axis transformation to find the solution the eigenwert and the proper vector of arbitrary shape object covariance, side information---the ellipse of inertia of structure object at the watermark built-in end; Utilize the image scrambling technology with significant watermark information scramble then, and utilize the form adaptive cosine transform, and, the watermark information behind the scramble is embedded in the object of arbitrary shape in conjunction with the human vision shadow model; By the inverse-DCT conversion, obtain the object of embed watermark at last.At the watermark detection end, at first utilize ellipse of inertia technology to obtain by the motor synchronizing information of object of attack, by with the comparison of primary object side information, realize the quick motor synchronizing calibration of object; Utilize Wiener filtering to predict the not object data of embed watermark then,, realize the detection of watermark with under the situation that does not have original image to participate in; The watermark information that extracts recovers through scramble, carry out correlativity with original watermark information and judge, by with the comparison of certain threshold level, judge whether the precognition watermark exists.
Fig. 2 is the video test sequence legend.
Fig. 2 (a) is the 3rd frame of Foreman video sequence; 2 (b) cover for the object two-value that this frame is partitioned into; 2 (c) are the corresponding object ellipse of inertia; 2 (d) are watermark information; 2 (e) are the frame of video of embed watermark; 2 (f) are under the situation that does not have to attack, the watermark information that extracts.
Fig. 3 is the form adaptive DCT coding diagram of arbitrary shape object.
Fig. 3 (a) is the arbitrary shape object; 3 (b) are the sub-piece in object one border; 3 (c) arrange sub-piece pixel by the coboundary, carry out vertical dct transform; Positioning DC coefficient behind the vertical DCT of 3 (d); 3 (e) arrange sub-piece pixel by left margin, carry out horizontal dct transform; 3 (f) positioning DC coefficient DC.
Fig. 4 is the motor synchronizing detectability analysis diagram of the ellipse of inertia.
Fig. 4 (a) detects error for ellipse of inertia scaling; 4 (b) detect error for the ellipse of inertia anglec of rotation;
Fig. 5 attacks for convergent-divergent and the performance evaluation diagram of water mark method is down attacked in rotation.
Fig. 5 (a) attacks watermark detection response diagram down for convergent-divergent; 5 (b) attack watermark detection response diagram down for rotation.
Embodiment:
For understanding technical scheme of the present invention better, describe in further detail below in conjunction with drawings and Examples.
1, object edges information obtains
Before digital watermarking embeds, at first the object of embed watermark is not carried out ellipse of inertia analysis, realize the motor synchronizing watermark detection with the side information that obtains object.The present invention utilizes the principal axis transformation in the pattern-recognition to find, requires the value of object major axes orientation z, can change into and ask object covariance matrix C
xThe proper vector problem.The barycentric coordinates of arbitrary shape object can be according to the calculating of having a few of subject area:
Center of gravity with object is a true origin, and then its covariance matrix can be expressed as:
Wherein, I
XImage is defined as the object rigid body around X-axis moment of inertia, I
YFor around the Y-axis moment of inertia, reach I
XYBe product of inertia.According to covariance matrix C
xSecular equation, find the solution and obtain its eigenvalue
k(k=1,2) are:
Thus, 2 the half long a of main shaft and the b of the definition ellipse of inertia are respectively:
According to covariance matrix C
xProper vector υ
k(k=1,2), the inclination angle phi of 2 main shafts of the definition ellipse of inertia
1And φ
2(being ellipse of inertia main shaft and the folded acute angle of X-axis forward) is respectively:
So far, the side information of arbitrary shape object---the ellipse of inertia can be definite fully according to the direction and the length of the barycentric coordinates of trying to achieve object, two main shafts.The object in difformity, size and orientation has different ellipse of inertia characteristic parameters, and when object suffered geometric attack, characteristic parameter also can change thereupon.Therefore, can characterize the object of an arbitrary shape and size with the ellipse of inertia: oval area approximation is in the area of primary object, the center of gravity of the corresponding primary object in oval center, the length of two main shafts and direction can provide convergent-divergent, the rotation information of object respectively.Even object suffers geometric attacks such as displacement, rotation, convergent-divergent, still can adjust object and detect to obtain watermark synchronization.The benefit of the ellipse of inertia is that obtaining of synchronizing information is independent of the position of object in frame of video, just in time satisfies the requirement of MPEG-4.
2, the embedding of digital watermarking
With the Foreman video sequence is example, and the embedding and the testing process of digital watermarking is described.Fig. 2 (a) has shown the 3rd frame of Foreman video sequence, and Fig. 2 (b) and Fig. 2 (c) are respectively the object ellipse of inertia that object two-value that this frame is partitioned into is covered and calculated.Fig. 2 (d) is a watermark information, and it is a significant bianry image, and size is 80 * 68.Consider the tolerance of human eye, distribute, and the bit that is damaged when disperseing watermark detection for making watermark signal more approach random noise for random noise, adopt the Arnold conversion with the watermark information scramble after, carry out watermark again and embed.
It is the linear adder strategy that watermark embeds rule, and watermark signal is hidden in the special frequency channel of object according to the shape and the vision shadow model of object.At first the object of arbitrary shape is divided by 8 * 8 sub-pieces; And, image block is divided into background piece and object piece, and the object piece can be divided into object home block and object bounds piece, shown in Fig. 3 (a) according to the shape information of object.The object home block is encoded with standard DCT, and the object bounds piece realizes that with form adaptive DCT coding watermark embeds.Any boundary block with object is example (as Fig. 3 (b)), and the implementation of form adaptive DCT is as follows: at first with each row coefficient of image block by coboundary alignment, and, carry out 1 dimension dct transform at vertical direction, shown in Fig. 3 (c) according to the length-specific of each row; Then, the coefficient after the conversion is pressed the left margin alignment, carry out 1 dimension dct transform of length-specific in the horizontal direction, shown in Fig. 3 (e); At last, the DC coefficient is positioned the upper left corner of 8 * 8 image blocks, shown in Fig. 3 (f).Because human eye is responsive to the noise ratio of low-frequency range, and high band information might be lost in compression process, so watermark information should embed the Mid Frequency of DCT coefficient, to obtain the balance of robustness and invisibility.
The dct transform coefficient of 8 * 8 sub-pieces is pressed ZigZag scanning arrange, the coefficient from (L+1) th to (L+N) th is used for embed watermark.It is as follows to embed rule:
I
i′＝I
l+α
l·w
i (9)
I in the formula
l', I
lBe respectively DCT coefficient and the original DCT coefficient behind the image embed watermark, w
iIt is watermark signal.α
lBe the weighting factor that watermark embeds, by human vision shadow model adaptive determining, can be under sightless prerequisite, the embed watermark of maximum intensity.For the sub-piece in object inside, the intermediate frequency coefficient scope that watermark embeds is selected L=13, N=17; And for the sub-piece in border, the intermediate frequency range that watermark embeds is selected L=9, N=5; Do not embed if the sub-piece in border has enough coefficients to be used for watermark, will give it up.At last,, carry out reverse SA-DCT, obtain the object of embed watermark according to the shape information of object.The frame of video of embed watermark is shown in Fig. 2 (e), and Y-PSNR is PSNR=46.509.
3, the motor synchronizing of object calibration
Before digital watermarking detects, at first carry out being calibrated by the motor synchronizing of object of attack.Utilize the ellipse of inertia technical Analysis information such as center of gravity, major axes orientation and length of object under fire, and the object edges information before embedding with watermark compares, to adjust the synchronism that object obtains watermark detection.It is pointed out that the center of gravity of destination object should be complementary with the center of its ellipse of inertia in the process of object calibration, not only, also divide for the image block of controlling object in the watermark extraction process in order to carry out accurate object calibration.
In order to estimate the ability of ellipse of inertia forecasting object motor synchronizing information, carry out the l-G simulation test that convergent-divergent and rotation are attacked.At first extract the side information of arbitrary shape object with ellipse of inertia technology, respectively object video is carried out 20%～200% X, the Y-axis object rotary manipulation of zoom operations and-90 °～90 ° simultaneously then, utilize ellipse of inertia technology to come the parameter of geometric attack that forecasting object suffers again.Fig. 4 has shown the test findings of ellipse of inertia Performance Detection, Fig. 4 (a) show ellipse of inertia scaling detect error (the object scaling of ellipse of inertia prediction and actual scaling poor) and be no more than ± 5.0 * 10
-3Fig. 4 (b) shows that the ellipse of inertia anglec of rotation detects error (the object anglec of rotation of ellipse of inertia prediction and the actual anglec of rotation poor) and is no more than+0.2 °, because it is very little to detect error, thereby proved the ellipse of inertia under single geometric attack, the validity of prediction motor synchronizing information.It is to be noted, because the ellipse of inertia is to try to achieve the anglec of rotation according to the slope of characteristics of objects vector, therefore, the accurately rotation of detected object in-90 °～+ 90 ° scopes only, if exceed above scope, the anglec of rotation that the ellipse of inertia obtains will differ 180 ° with the actual anglec of rotation.In order to improve the watermark detection rate, in actual applications, can carry out watermark detection by two angles difference calibration object.
4, the detection of digital watermarking
The testing process of digital watermarking is the inverse process of digital watermark embed process basically.For realizing that watermark detection does not need original image, utilize Wiener filtering to come the raw value of forecasting object.Embed the zone in watermark, the DCT value I ' of destination object (i) has determined the watermark predicted value s (i) at this place with the DCT value I ' difference symbol (i) of forecasting object:
s(i)＝sign(I′(i)- I′(i)) (10)
After the watermark s (i) that extracts recovers through the Arnold scramble and original watermark signal w (i) be calculated as follows similarity, the quality of the watermark information that obtains with the objective evaluation algorithm:
If (s w) greater than predetermined watermark detection threshold value T, is that watermark exists with regard to decidable to similarity Sim.Under the situation that is not having to attack, from the watermark information that extracts the object shown in Fig. 2 (f), watermark detection similarity Sim=0.938.
In order to estimate the ability that water mark method opposing convergent-divergent is attacked, the object video of embed watermark is carried out X, the Y-axis challenge trial of convergent-divergent in proportion, the scope 30%～600% of scaling.Fig. 5 (a) attacks watermark detection response diagram down for convergent-divergent, can see, water mark method is relatively more responsive to dwindling attack, and the watermark detection similarity reduces gradually along with reducing gradually of scaling; But to amplification attack, the watermark detection rate changes more steady.This is because reduction operation can make image lose partial information and can't recover, thereby has influence on the watermark detection precision.In order to estimate the robustness that water mark method is attacked rotation, the object video of embed watermark carried out-90 °～+ 90 ° rotation challenge trial, Fig. 5 (b) attacks watermark detection response diagram down for rotation, and horizontal ordinate is the object anglec of rotation, and ordinate is the watermark detection similarity.Can see that under the rotation of different angles was attacked, the watermark detection similarity maintained about 0.81 substantially, changes steadily.
Claims (1)
1, OO motor synchronizing method of detecting watermarks in a kind of media product is characterized in that comprising following concrete steps:
1) obtaining of object edges information: utilize the principal axis transformation in the pattern-recognition, structure is used for describing side information---the ellipse of inertia of arbitrary shape object orientation and size information, before digital watermarking embeds, at first the object of embed watermark is not carried out ellipse of inertia analysis, eigenwert and proper vector according to the object covariance matrix, obtain the center of gravity of object, the direction and the big or small side information of main shaft, and be delivered to the motor synchronizing calibration that the watermark detection end is used for object;
2) embedding of digital watermarking: adopt the Arnold conversion with the watermark scramble, carrying out watermark again embeds, it is the linear adder strategy that watermark embeds rule, the object of arbitrary shape is divided by 8 * 8 sub-pieces, and according to the shape information of object, image block is divided into background piece and object piece, the object piece is divided into object home block and object bounds piece again, the object home block is encoded with the normal scatter cosine transform, and object bounds piece form adaptive discrete cosine transform coding realizes that watermark embeds, watermark information embeds the Mid Frequency of discrete cosine transform coefficient, according to the shape information of object, carry out reverse discrete cosine transform at last, obtain the object of embed watermark;
3) motor synchronizing of object calibration: in the testing process of digital watermarking, the motor synchronizing information of at first carrying out object detects, utilize the ellipse of inertia to analyze under fire center of gravity, major axes orientation and the length information of object, and the object edges information before embedding with watermark compares, to adjust the synchronism that object obtains watermark detection, in the object calibration process, the center of gravity of destination object should be complementary with the center of its ellipse of inertia;
4) detection of digital watermarking: utilize Wiener filtering to come the raw value of forecasting object, embed the zone in watermark, the difference symbol of the discrete cosine transform coefficient value of destination object and the discrete cosine transform coefficient value of forecasting object has determined the watermark predicted value at this place, after the watermark that dopes recovers through the Arnold scramble, the correlativity of predicting watermark and original watermark compares, if similarity greater than the preset detection threshold value, is that watermark exists with regard to decidable.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 03151038 CN1241389C (en) | 2003-09-18 | 2003-09-18 | Object-oriented self-synchronizing water mark detecting method for multimedia product |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 03151038 CN1241389C (en) | 2003-09-18 | 2003-09-18 | Object-oriented self-synchronizing water mark detecting method for multimedia product |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN1527257A true CN1527257A (en) | 2004-09-08 |
| CN1241389C CN1241389C (en) | 2006-02-08 |
Family
ID=34286867
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN 03151038 Expired - Fee Related CN1241389C (en) | 2003-09-18 | 2003-09-18 | Object-oriented self-synchronizing water mark detecting method for multimedia product |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN1241389C (en) |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1322470C (en) * | 2005-07-04 | 2007-06-20 | 杭州天谷信息科技有限公司 | Digital watermark anti fake method in anti-digital-to-analog conversion process |
| CN100357971C (en) * | 2006-01-18 | 2007-12-26 | 李京兵 | Wavelet-based geometric attack resistant digital watermark method |
| CN101437163A (en) * | 2009-01-08 | 2009-05-20 | 中国科学院研究生院 | Multi-watermark technology based on network information theory |
| CN101841700A (en) * | 2010-05-18 | 2010-09-22 | 宁波大学 | H.264 compression bit stream oriented video watermark method |
| CN101888548A (en) * | 2010-06-17 | 2010-11-17 | 宁波大学 | H.264/AVC video information content authentication-based fragile watermarking method |
| CN102024248A (en) * | 2010-11-30 | 2011-04-20 | 辽宁师范大学 | Digital image watermarking method based on local visual attention |
| CN102156952A (en) * | 2011-03-08 | 2011-08-17 | 南京师范大学 | Copyright-protection-service-oriented remote sensing image digital watermarking method |
| CN101308504B (en) * | 2007-05-16 | 2013-04-10 | 佳能株式会社 | Image processing apparatus and image retrieval method |
| CN102132564B (en) * | 2008-08-20 | 2013-09-11 | 汤姆森特许公司 | Selection of watermarks for the watermarking of compressed video |
| CN103854250A (en) * | 2014-03-17 | 2014-06-11 | 广东石油化工学院 | Self-synchronizing image watermarking method applied to wireless multimedia sensor network |
| CN108154212A (en) * | 2017-12-13 | 2018-06-12 | 上海秦镜网络科技有限公司 | A kind of coding and decoding methods based on horizontal and vertical parity check code |
| CN108335254A (en) * | 2017-01-17 | 2018-07-27 | 腾讯科技（深圳）有限公司 | A kind of data information processing method and device |
| CN113393363A (en) * | 2021-06-29 | 2021-09-14 | 合肥高维数据技术有限公司 | Watermark embedding method, watermark extracting method, storage medium and electronic equipment |
-
2003
- 2003-09-18 CN CN 03151038 patent/CN1241389C/en not_active Expired - Fee Related
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1322470C (en) * | 2005-07-04 | 2007-06-20 | 杭州天谷信息科技有限公司 | Digital watermark anti fake method in anti-digital-to-analog conversion process |
| CN100357971C (en) * | 2006-01-18 | 2007-12-26 | 李京兵 | Wavelet-based geometric attack resistant digital watermark method |
| CN101308504B (en) * | 2007-05-16 | 2013-04-10 | 佳能株式会社 | Image processing apparatus and image retrieval method |
| CN102132564B (en) * | 2008-08-20 | 2013-09-11 | 汤姆森特许公司 | Selection of watermarks for the watermarking of compressed video |
| CN101437163A (en) * | 2009-01-08 | 2009-05-20 | 中国科学院研究生院 | Multi-watermark technology based on network information theory |
| CN101437163B (en) * | 2009-01-08 | 2013-05-29 | 中国科学院研究生院 | Multi-watermark technology based on network information theory |
| CN101841700A (en) * | 2010-05-18 | 2010-09-22 | 宁波大学 | H.264 compression bit stream oriented video watermark method |
| CN101888548A (en) * | 2010-06-17 | 2010-11-17 | 宁波大学 | H.264/AVC video information content authentication-based fragile watermarking method |
| CN102024248A (en) * | 2010-11-30 | 2011-04-20 | 辽宁师范大学 | Digital image watermarking method based on local visual attention |
| CN102156952B (en) * | 2011-03-08 | 2012-08-22 | 南京师范大学 | Copyright-protection-service-oriented remote sensing image digital watermarking method |
| CN102156952A (en) * | 2011-03-08 | 2011-08-17 | 南京师范大学 | Copyright-protection-service-oriented remote sensing image digital watermarking method |
| CN103854250A (en) * | 2014-03-17 | 2014-06-11 | 广东石油化工学院 | Self-synchronizing image watermarking method applied to wireless multimedia sensor network |
| CN108335254A (en) * | 2017-01-17 | 2018-07-27 | 腾讯科技（深圳）有限公司 | A kind of data information processing method and device |
| CN108154212A (en) * | 2017-12-13 | 2018-06-12 | 上海秦镜网络科技有限公司 | A kind of coding and decoding methods based on horizontal and vertical parity check code |
| CN113393363A (en) * | 2021-06-29 | 2021-09-14 | 合肥高维数据技术有限公司 | Watermark embedding method, watermark extracting method, storage medium and electronic equipment |
| CN113393363B (en) * | 2021-06-29 | 2024-05-07 | 合肥高维数据技术有限公司 | Watermark embedding and extracting method, storage medium and electronic device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN1241389C (en) | 2006-02-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Luo et al. | Edge adaptive image steganography based on LSB matching revisited | |
| CN101699508B (en) | Image digital watermark embedding and extracting method and system | |
| Bhatnagar et al. | A new robust reference watermarking scheme based on DWT-SVD | |
| Huan et al. | Exploring stable coefficients on joint sub-bands for robust video watermarking in DT CWT domain | |
| CN1241389C (en) | Object-oriented self-synchronizing water mark detecting method for multimedia product | |
| Peng et al. | A complete passive blind image copy-move forensics scheme based on compound statistics features | |
| CN102917227B (en) | Compressive sensing-based adaptive video information hiding method | |
| CN101908201B (en) | Robust digital image adaptive zero-watermarking method | |
| Patil et al. | DWT based invisible watermarking technique for digital images | |
| CN102223561B (en) | Blind watermark embedding and extracting method of stereoscopic video image | |
| Khalifa et al. | A robust non-blind algorithm for watermarking color images using multi-resolution wavelet decomposition | |
| CN1276387C (en) | Synchronous self-adaptable watermark method based on image continuity | |
| CN1738353A (en) | Digital watermark technology for resisting rotary extension and displacement attack | |
| CN102065287B (en) | T-shaped video watermark embedding and extracting method | |
| Ali et al. | Copyright protection for digital image by watermarking technique | |
| Wen et al. | An Adaptive JPEG Steganographic Scheme Based on Run‐Length Statistical Complexity | |
| CN109410115B (en) | Adaptive capacity image blind watermark embedding and extracting method based on SIFT feature points | |
| Zhi et al. | Edge detection based JND model for digital watermarking | |
| Kim et al. | An object-based video watermarking | |
| Liang et al. | Ridgelet-based robust and perceptual watermarking for images | |
| Jie et al. | Object-based watermarking scheme robust to object manipulations | |
| Xu et al. | Estimation of MB steganography based on least square method | |
| Wang | Research on a Watermarking Algorithm for Remote Sensing Images Resist Screen Capture Attacks in Urban Planning Information Management | |
| Chauhan | Information hiding watermarking detection technique by psnr and rgb intensity | |
| Chen et al. | A digital watermarking algorithm based on characters of the remote-sensing imagery |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| C19 | Lapse of patent right due to non-payment of the annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |