CN110163787A - Digital audio Robust Blind Watermarking Scheme embedding grammar based on dual-tree complex wavelet transform - Google Patents
Digital audio Robust Blind Watermarking Scheme embedding grammar based on dual-tree complex wavelet transform Download PDFInfo
- Publication number
- CN110163787A CN110163787A CN201910343271.XA CN201910343271A CN110163787A CN 110163787 A CN110163787 A CN 110163787A CN 201910343271 A CN201910343271 A CN 201910343271A CN 110163787 A CN110163787 A CN 110163787A
- Authority
- CN
- China
- Prior art keywords
- information
- watermark
- signal
- embedded
- audio
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T1/00—General purpose image data processing
- G06T1/0021—Image watermarking
- G06T1/005—Robust watermarking, e.g. average attack or collusion attack resistant
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2201/00—General purpose image data processing
- G06T2201/005—Image watermarking
- G06T2201/0052—Embedding of the watermark in the frequency domain
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2201/00—General purpose image data processing
- G06T2201/005—Image watermarking
- G06T2201/0061—Embedding of the watermark in each block of the image, e.g. segmented watermarking
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2201/00—General purpose image data processing
- G06T2201/005—Image watermarking
- G06T2201/0065—Extraction of an embedded watermark; Reliable detection
Abstract
The invention discloses a kind of digital audio Robust Blind Watermarking Scheme embedding grammar based on dual-tree complex wavelet transform, using the source code flow of audio signal as the embedded object of watermark information, correlation extremely low between pseudo-random sequence and natural audio signal is utilized, has fully ensured that a possibility that embedding information is correctly extracted；Watermark information by Infinite Cyclic is embedded into whole section audio signal, to greatly promote extraction accuracy rate, and algorithm is made to obtain cutting robustness, the audio after insertion and former audio are a kind of non digital blind watermark embedding grammars of the sense of hearing in acoustically unaware difference.Watermark is embedded in the dual-tree complex wavelet transform domain of image; the method being embedded in using piecemeal, is increased information embedded quantity, can resist the common attacks such as image quality compression, the transformation of scale of image; with certain robustness, there is very great meaning to the copyright protection of image or audio.
Description
Technical field
The present invention proposes a kind of image digital watermark embedding grammar, and in particular to the audio number based on dual-tree complex wavelet transform
Word Robust Blind Watermarking Scheme embedding grammar；Belong to image and technical field of video processing.
Background technique
It is same in multimedia with the rapid development of mobile network in recent years, it is more including digital picture, digital audio
The propagation of media product becomes increasingly easy, and spread speed is getting faster；And the duplication of multimedia content becomes increasingly simpler
Single, pirate difficulty is lower and lower, seriously compromises the legitimate rights and interests of creator.Thus, same Copyright Protection in multimedia
Increasingly severe, audio watermarking has begun and one of has become a hot topic of research as a kind of feasible scheme.
Audio watermarking is built-in the secret signal in audio signal, it allow audio creator can to oneself
Audio works establishes copyright authentication, marks the owner, is a kind of reliable technological means for realizing multimedia content copyright protection.
With the development of audio-frequency signal coding technology, the development of audio watermarking technique is also to weed out the old and bring forth the new.The visible logo of early application
It as audio frequency watermark, is directly placed in a region of audio, such method is easy to use, but such watermark is for carrying
The quality of body audio is affected, and causes sound quality not good enough, and is easy cropped.Recently, some research achievements propose a series of
Robust can not perceive Audio Watermarking Algorithm, these algorithms generally have a very strong robustness or not sentience, but sound quality with
Robustness, conflict when as watermark signal insertion, cannot access good balance always.
In view of the foregoing, it is necessary to a kind of new embedding grammar is proposed, to take into account the robustness and carrier sound of watermark
The sound quality of frequency.
Summary of the invention
To solve the deficiencies in the prior art, it is an object of the invention to propose a kind of audio based on dual-tree complex wavelet transform
Digital robust blind watermatking embedding grammar, algorithm insertion is one section of sightless digital information, and is dispersed in each frame of audio
Frequency domain among, a variety of common audios attacks can be effective against, can preferably guarantee the quality containing watermarked audio, thus
Take into account the robustness of watermark and the sound quality of carrier audio.
In order to achieve the above objectives, the present invention adopts the following technical scheme that:
Using the source code flow of audio signal as the embedded object of watermark information, pass through the design of algorithm, it is ensured that system is anti-
Attack robust is sufficiently high, meanwhile, and carrier signal sound quality can be effectively ensured.
Firstly, reading carrier signal audio sample point information, and all sampled points are subjected to sub-frame processing.The present invention utilizes
Correlation extremely low between pseudo-random sequence and natural audio signal, has fully ensured that the possibility that embedding information is correctly extracted
Property.Watermark is spliced by multiple pseudo-random sequences, these pseudo-random sequences are generated by 16 different keys, to these
0-15, corresponding 16 binary information 0x0-F, so each pseudo-random sequence represents the letter of 4bit is numbered in basic watermark
Breath.To each frame, pseudo-random sequence is embedded into carrier audio by we, obtains a big watermark, can thus obtain
To biggish embedding capacity.2 layers DTCWT (dual-tree complex wavelet transform) frequency domain conversion process is done to pseudo-random sequence to be embedded
Afterwards, it is embedded into carrier audio.In order to guarantee that watermark information can be extracted correctly, the application is advised using the insertion of " two repetitions "
Then, i.e., each code element repeats in insertion multiframe content, meanwhile, watermark information Infinite Cyclic it is embedded into whole section audio signal,
To greatly promote extraction accuracy rate, and algorithm is made to obtain cutting robustness.
After being embedded in watermark information, then the time-domain signal being embedded in for realizing frame synchronization, and audio is saved as aqueous official seal
Number.Audio and former audio after insertion is a kind of non digital blind watermark insertion of the sense of hearing in acoustically unaware difference
Method.When extraction, the signal received and synchronous time domain signal are first done into cross-correlation calculation, obtain the position of synchronization signal, it is real
Existing frame synchronization then does sub-frame processing to audio, and to each frame, we are converted with 2 layers of DTCWT takes out intermediate frequency subband, and makes
The pseudo-random sequence of itself and generation calculates correlation, thus the watermark sequence being embedded in.
Detailed process is as follows for watermark insertion:
(1) audio framing and sequence map
Firstly, carrier audio cutting is several frames by information insertion person, each frame includes 1024 time-domain samplings of sequence
Point, end less than a frame sampled point then not within the scope of insertion.Later, 16 different pseudo-random seeds are selected, are used to
16 different pseudo-random sequences are generated, the value range of element is [- 1,1], the range and sampled audio signal point in sequence
Amplitude it is consistent.In this algorithm, it is embedded in different watermarks, that is, represents and is embedded in different hexadecimal watermark informations.This 16
A pseudo-random sequence is as sequence of mapping, number consecutively 0-15, corresponding hexadecimal 0x0F.Therefore, each pseudo-random sequence
Insertion is embedded in corresponding to the binary system watermark information of 4bit.Therefore, watermark information is mapped as sequence to be embedded by information insertion person.
In addition, information insertion person selects the 17th pseudo-random seed, for generating audio synchronous signal, audio sync is in step (3)
It introduces.
(2) sequence of mapping is embedded in
Information insertion person is by superimposed by the DTCWT intermediate-freuqncy signal of sequence to be embedded and original audio signal, realization water
The insertion of official seal breath.Firstly, for each frame, insertion person does 2 layers of dual-tree complex wavelet to sequence to be embedded and original series respectively
It converts (2-level DTCWT), by transformation, signal is switched into low frequency, intermediate frequency and high-frequency sub-band by time domain.In order to guarantee information
The influence being embedded in original signal is as small as possible, while being also required to possess enough robustness, chooses intermediate frequency subband as information
The target area of insertion.Assuming that the second layer DTCWT transformation intermediate frequency coefficient difference of original signal, sequence to be embedded and output sequence
For A, X, W, then the insertion of sequence of mapping meets following rule:
W=A+2X
α=E (X)/E (A)=10*log (SNR) (a)
In formula (a), α is the embedment strength of signal, and E () is the energy of signal, and SNR is the signal-to-noise ratio containing watermark signal.One
As, it is demonstrated experimentally that insertion watermark information sense of hearing caused by original signal influences basic when SNR value is greater than 60 (dB)
Very little can satisfy " the not sentience " of watermark, while it is also ensured that watermark correct extraction.Frequency domain line more than
Property superposition after, being inverse DTCWT to W converts, so that it may obtain output audio signal with watermarked information.
In order to guarantee that the robustness of embedding information, system define each code word needs and continuously repeat insertion 4 times, Ye Jilian
The code word being embedded in continuous 4 content frames be it is the same, overall structure is as shown in Figure 3.In this way, even if may be sent out once in a while extracting end
Raw code word extracts error, but system can take out the wherein maximum code of statistical probability after the code word by continuously extracting 4 frames
Word extracts code word as correct.Meanwhile watermark information will by Infinite Cyclic be embedded into entire carrier audio signal.Cause
This, even if original audio is cut, information extraction person also have very maximum probability can at least be observed from remaining audio it is whole
The information insertion of section.
(3) synchronizing information is embedded in
After with watermarked information cut, user needs to be accurately positioned the initial position of each frame of audio signal, with reality
Existing signal is synchronous, just may insure the accuracy of information extraction.In addition, recipient is also required to know that complete watermark information starts
The position of insertion.Therefore, it after information insertion person completes the insertion of watermark every time, is all completed in a back to back frame (synchronization frame)
The pseudo-random signal of time domain is superimposed, and superposition rule here is substantially the same with formula (a), different from, in order to guarantee to mention
The accuracy positioned when taking synchronizing information to each frame boundaries, superposition here are completed in the time domain, namely are completed complete
Frequency is superimposed.
Detailed process is as follows for watermark extracting:
(1) pseudo-random sequence generates
Information extraction person needs to share identical pseudo-random seed with information receiver, to generate identical pseudorandom sequence
Column, and the pseudo-random sequence for being used for hiding information is done 2-level DTCWT transformation, and retain intermediate frequency subband, by high frequency with it is low
Frequency subband is not counted the small change, then the signal for only being remained intermediate frequency is converted by inverse DTCWT, is abbreviated asPuppet with
The shared of machine can be realized by arranging in advance or in the case where more secret communication mode is transmitted.Equally, pseudo-random seed can also
To be owned as key by information insertion person, and it is supplied to legal information receiver, realizes better information encryption.
(2) signal is synchronous
After information extraction person receives audio signal, relativity of time domain is carried out using synchronizing sequence and whole section audio signal first
It calculates, and obtains a correlation peaks, it is assumed that be Cp, for remaining sampled point, if the correlation in the point is greater than
Or it is equal to T*Cp, then think that the sampled point is also the initial position of synchronization frame.Here T is a decision threshold, default setting
It is 0.8, higher thresholding more has the ability for excluding synchronous interference, but can also generate higher synchronization loss probability, lower
Thresholding may then introduce higher possible false synchronous.
(3) correlation calculations
After signal synchronizes, recipient carries out 2-level DTCWT transformation to each frame signal, and retains intermediate frequency subband, will
High frequency is not counted the small change with low frequency sub-band, then the signal for only being remained intermediate frequency is converted by inverse DTCWT, is abbreviated as WM.Then, recipient counts
Calculate WMWith PMThe correlation of middle element, and the corresponding pseudo-random sequence of correlation peaks is taken, the serial number of seed is that present frame mentions
The watermark information (M) obtained.
In formula (b), Cov (x, y) indicates to calculate the correlation of x and y.
(4) statistical average result
The highest symbol of the frequency of occurrences is taken by the watermark information that continuous 4 frame extracts, we can accurately obtain embedding
The watermark information entered.Also, if extractor has obtained more than one synchronous code in audio, for each symbol,
The frequency of result can also be extracted in different circulations according to them, takes out the highest symbol of frequency of occurrence as a result, therefore
Information extraction accuracy rate can further be promoted.
The invention has the beneficial effects that:
(1) embedded object of the application using the source code flow of audio signal as watermark information is become based on dual-tree complex wavelet
Change the design by algorithm, it is ensured that system attack resistance robustness is sufficiently high, meanwhile, and carrier signal sound quality can be effectively ensured；
(2) correlation extremely low between pseudo-random sequence and natural audio signal is utilized, has fully ensured that embedding information
A possibility that correctly extraction.Moreover, pseudo-random sequence is embedded into carrier audio by we to each frame, one is obtained
Big watermark can be obtained by biggish embedding capacity in this way；
(3) watermark information by Infinite Cyclic is embedded into whole section audio signal, so that extraction accuracy rate is greatly promoted, and
So that algorithm obtains cutting robustness, the audio after insertion and former audio in acoustically unaware difference, be a kind of sense of hearing not
Appreciable digital blind watermark embedding grammar；
(4) by the embedding grammar of the application, a certain amount of information can be embedded in piece image or audio-video, such as
Image sequence number or some other copyright information.Watermark is embedded in the dual-tree complex wavelet transform domain of image, using piecemeal
The method of insertion increases information embedded quantity, can resist the common attacks such as image quality compression, the transformation of scale of image, has one
Fixed robustness has very great meaning to the copyright protection of image or audio.
Detailed description of the invention
Fig. 1 is the frame of watermark telescopiny in the total algorithm frame of audio Robust Blind Watermarking Scheme embedding grammar of the invention
Figure；
Fig. 2 is the frame of watermark extraction process in the total algorithm frame of audio Robust Blind Watermarking Scheme embedding grammar of the invention
Figure；
Fig. 3 is that watermark telescopiny sound intermediate frequency framing and sequence map schematic diagram；
Fig. 4 is that watermark signal and synchronization signal are embedded in schematic diagram in watermark telescopiny.
Specific embodiment
Specific introduce is made to the present invention below in conjunction with the drawings and specific embodiments.
Digital audio Robust Blind Watermarking Scheme embedding grammar based on dual-tree complex wavelet transform of the invention particularly may be divided into watermark
Insertion and watermark extracting two parts.
(1) frame diagram of watermark telescopiny is referring to Fig. 1:
(1) audio framing and sequence map
The frame diagram of audio framing and sequence mapping is referring to Fig. 3.Firstly, carrier audio cutting is several by information insertion person
Frame, each frame include 1024 time domain sampling points of sequence, end less than a frame sampled point then not within the scope of insertion.It
Afterwards, 16 different pseudo-random seeds, for generating 16 different pseudo-random sequences, the value range of element in sequence are selected
For [- 1,1], the range is consistent with the amplitude of sampled audio signal point.In this algorithm, it is embedded in different watermarks, that is, is represented
It is embedded in different hexadecimal watermark informations.This 16 pseudo-random sequences correspond to 16 as sequence of mapping, number consecutively 0-15
System 0x0F.Therefore, the insertion of each pseudo-random sequence is embedded in corresponding to the binary system watermark information of 4bit.Therefore, information is embedding
Watermark information is mapped as sequence to be embedded by the person of entering.In addition, information insertion person selects the 17th pseudo-random seed, for generating sound
Vertical synchronizing signal, audio sync are introduced in step (3).
(2) sequence of mapping is embedded in
Information insertion person is by superimposed by the DTCWT intermediate-freuqncy signal of sequence to be embedded and original audio signal, realization water
The insertion of official seal breath.Firstly, for each frame, insertion person does 2 layers of dual-tree complex wavelet to sequence to be embedded and original series respectively
It converts (2-level DTCWT), by transformation, signal is switched into low frequency, intermediate frequency and high-frequency sub-band by time domain.In order to guarantee information
The influence being embedded in original signal is as small as possible, while being also required to possess enough robustness, chooses intermediate frequency subband as information
The target area of insertion.Assuming that the second layer DTCWT transformation intermediate frequency coefficient difference of original signal, sequence to be embedded and output sequence
For A, X, W, then the insertion of sequence of mapping meets following rule:
W=A+ α X
α=E (X)/E (A)=10*log (SNR) (a)
In formula (a), α is the embedment strength of signal, and E () is the energy of signal, and SNR is the signal-to-noise ratio containing watermark signal.One
As, it is demonstrated experimentally that insertion watermark information sense of hearing caused by original signal influences basic when SNR value is greater than 60 (dB)
Very little can satisfy " the not sentience " of watermark, while it is also ensured that watermark correct extraction.Frequency domain line more than
Property superposition after, being inverse DTCWT to W converts, so that it may obtain output audio signal with watermarked information.
In order to guarantee that the robustness of embedding information, system define each code word needs and continuously repeat insertion 4 times, Ye Jilian
The code word being embedded in continuous 4 content frames be it is the same, overall structure is as shown in Figure 4.In this way, even if may be sent out once in a while extracting end
Raw code word extracts error, but system can take out the wherein maximum code of statistical probability after the code word by continuously extracting 4 frames
Word extracts code word as correct.Meanwhile watermark information will by Infinite Cyclic be embedded into entire carrier audio signal.Cause
This, even if original audio is cut, information extraction person also have very maximum probability can at least be observed from remaining audio it is whole
The information insertion of section.
(3) synchronizing information is embedded in
After with watermarked information cut, user needs to be accurately positioned the initial position of each frame of audio signal, with reality
Existing signal is synchronous, just may insure the accuracy of information extraction.In addition, recipient is also required to know that complete watermark information starts
The position of insertion.Therefore, it after information insertion person completes the insertion of watermark every time, is all completed in a back to back frame (synchronization frame)
The pseudo-random signal of time domain is superimposed, and superposition rule here is substantially the same with (1), different from, same in order to guarantee to extract
The accuracy positioned when walking information to each frame boundaries, superposition here are completed in the time domain, namely are completed full range and folded
Add.
(2) frame diagram of watermark extraction process is referring to fig. 2:
(1) pseudo-random sequence generates
Information extraction person needs to share identical pseudo-random seed with information receiver, to generate identical pseudorandom sequence
Column, and the pseudo-random sequence for being used for hiding information is done 2-level DTCWT transformation, and retain intermediate frequency subband, by high frequency with it is low
Frequency subband is not counted the small change, then the signal for only being remained intermediate frequency is converted by inverse DTCWT, is abbreviated asPuppet with
The shared of machine can be realized by arranging in advance or in the case where more secret communication mode is transmitted.Equally, pseudo-random seed can also
To be owned as key by information insertion person, and it is supplied to legal information receiver, realizes better information encryption.
(2) signal is synchronous
After information extraction person is by audio signal, relativity of time domain is carried out using synchronizing sequence and whole section audio signal first
It calculates, and obtains a correlation peaks, it is assumed that be Cp, for remaining sampled point, if the correlation in the point is greater than
Or it is equal to T*Cp, then think that the sampled point is also the initial position of synchronization frame.Here T is a decision threshold, default setting
It is 0.8, higher thresholding more has the ability for excluding synchronous interference, but can also generate higher synchronization loss probability, lower
Thresholding may then introduce higher possible false synchronous.
(3) correlation calculations
After signal synchronizes, recipient carries out 2-level DTCWT transformation to each frame signal, and retains intermediate frequency subband, will
High frequency is not counted the small change with low frequency sub-band, then the signal for only being remained intermediate frequency is converted by inverse DTCWT, is abbreviated as WM.Then, recipient counts
Calculate WMWith PMThe correlation of middle element, and the corresponding pseudo-random sequence of correlation peaks is taken, the serial number of seed is that present frame mentions
The watermark information (M) obtained.
In formula (b), Cov (x, y) indicates to calculate the correlation of x and y.
(4) statistical average result
The highest symbol of the frequency of occurrences is taken by the watermark information that continuous 4 frame extracts, we can accurately obtain embedding
The watermark information entered.Also, if extractor has obtained more than one synchronous code in audio, for each symbol,
The frequency of result can also be extracted in different circulations according to them, takes out the highest symbol of frequency of occurrence as a result, therefore
Information extraction accuracy rate can further be promoted.
To sum up, by the embedding grammar of the application, a certain amount of information, example can be embedded in piece image or audio-video
Such as image sequence number or some other copyright information.Watermark is embedded in the dual-tree complex wavelet transform domain of image, using point
The method of block insertion, increases information embedded quantity, can resist the common attacks such as image quality compression, the transformation of scale of image, have
Certain robustness has very great meaning to the copyright protection of image or audio.
The basic principles, main features and advantages of the invention have been shown and described above.The technical staff of the industry should
Understand, the above embodiments do not limit the invention in any form, all obtained by the way of equivalent substitution or equivalent transformation
Technical solution is fallen within the scope of protection of the present invention.
Claims (10)
1. the image digitization piecemeal blind watermatking embedding grammar based on dual-tree complex wavelet transform, which is characterized in that including
(1) watermark is embedded in:
It (1) is several frames by carrier audio cutting, each frame includes 1024 time domain sampling points of sequence, and end is less than a frame
Sampled point is not then within the scope of insertion；
(2) 16 different pseudo-random seeds are selected, for generating 16 different pseudo-random sequences, and select the 17th puppet with
Machine, for generating audio synchronous signal；
(3) the DTCWT intermediate-freuqncy signal of sequence to be embedded is superimposed with original audio signal, realize the insertion of watermark information；Letter
Shown in the method such as formula (a) of number superposition, inverse DTCWT is done to W and is converted, output audio signal with watermarked information is obtained:
Assuming that the second layer DTCWT transformation intermediate frequency coefficient of original signal, sequence to be embedded and output sequence is respectively A, X, W, that
The insertion of sequence of mapping meets following rule:
W=A+ α X
α=E (X)/E (A)=10*log (SNR) (a)
In formula (a), α is the embedment strength of signal, and E () is the energy of signal, and SNR is the signal-to-noise ratio containing watermark signal；
(4) each code word need to continuously repeat insertion 4 times namely continuous 4 content frame in the code word that is embedded in be the same；
(5) after the completion of information insertion, the pseudo-random signal superposition of time domain, superposition rule are completed in a back to back frame (synchronization frame)
Then referring to formula (a), superposition here is completed in the time domain namely full range superposition；
(2) watermark extracting:
(1) information extraction person and information receiver use identical pseudo-random seed, will be used for the pseudo-random sequence of hiding information
2-level DTCWT transformation is done, and retains intermediate frequency subband, high frequency and low frequency sub-band are not counted the small change, then convert to obtain by inverse DTCWT
The signal for only remaining intermediate frequency, is abbreviated as
(2) relativity of time domain calculating is carried out using synchronizing sequence and whole section audio signal, and obtains a correlation peaks, it is false
It is set as Cp, for remaining sampled point, if the correlation in the point is greater than or equal to T*Cp, then think that the sampled point is also same
The initial position of step-frame, T are decision threshold；
(3) after signal synchronizes, recipient carries out 2-level DTCWT transformation to each frame signal, and retains intermediate frequency subband, will be high
Frequency is not counted the small change with low frequency sub-band, then the signal for only being remained intermediate frequency is converted by inverse DTCWT, is abbreviated as WM, then, recipient calculates
WMWith PMThe correlation of middle element, and the corresponding pseudo-random sequence of correlation peaks is taken, the serial number of seed is that present frame extracts
Obtained watermark information M；
(4) the highest symbol of the frequency of occurrences is taken by the watermark information that continuous 4 frame extracts, we can accurately obtain insertion
Watermark information.
2. the image digitization piecemeal blind watermatking embedding grammar according to claim 1 based on dual-tree complex wavelet transform, special
Sign is, the embedding method of " two repetitions " is used in this method, and each code element repeats in insertion multiframe content, meanwhile, watermark
Information Infinite Cyclic it is embedded into whole section audio signal.
3. the image digitization piecemeal blind watermatking embedding grammar according to claim 1 based on dual-tree complex wavelet transform, special
Sign is, using the source code flow of audio signal as the embedded object of watermark information.
4. the image digitization piecemeal blind watermatking embedding grammar according to claim 1 based on dual-tree complex wavelet transform, special
Sign is that 16 pseudo-random sequences are generated by 16 different keys, and 0-15 is numbered to these basic watermarks, corresponding
16 binary information 0x0-F, each pseudo-random sequence represent the information of 4bit.
5. the image digitization piecemeal blind watermatking embedding grammar according to claim 1 based on dual-tree complex wavelet transform, special
Sign is, in the step of (one) watermark is embedded in (3), for each frame, insertion person does 2 to insertion sequence and original series respectively
Signal is switched to low frequency, intermediate frequency and high-frequency sub-band by time domain by transformation by layer dual-tree complex wavelet transform (2-level DTCWT),
Choose the target area that intermediate frequency subband is embedded in as information.
6. the image digitization piecemeal blind watermatking embedding grammar according to claim 1 based on dual-tree complex wavelet transform, special
Sign is that SNR value is greater than 60dB.
7. the image digitization piecemeal blind watermatking embedding grammar according to claim 1 based on dual-tree complex wavelet transform, special
Sign is, (two) watermark extracting the step of in (1), pseudo-random seed it is shared by agreement in advance or in more secret communication side
Formula transmission is lower to be realized, can also be used as key is owned by information insertion person, and is supplied to legal information receiver, is realized more preferable
Information encryption.
8. the image digitization piecemeal blind watermatking embedding grammar according to claim 1 based on dual-tree complex wavelet transform, special
Sign is, (two) watermark extracting the step of in (2), decision threshold T the default setting is 0.8.
9. the image digitization piecemeal blind watermatking embedding grammar according to claim 1 based on dual-tree complex wavelet transform, special
Sign is, (two) watermark extracting the step of in (3), the expression publicity of watermark information M are as follows:
Wherein, Cov (x, y) indicates to calculate the correlation of x and y.
10. the image digitization piecemeal blind watermatking embedding grammar according to claim 1 based on dual-tree complex wavelet transform, special
Sign is, (two) watermark extracting the step of in (4), if extractor has obtained more than one synchronous code in audio, then right
In each symbol, the frequency of result is extracted in different circulations according to them, takes out the highest symbol of frequency of occurrence as knot
Fruit further promotes information extraction accuracy rate.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910343271.XA CN110163787B (en) | 2019-04-26 | 2019-04-26 | Audio digital robust blind watermark embedding method based on dual-tree complex wavelet transform |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910343271.XA CN110163787B (en) | 2019-04-26 | 2019-04-26 | Audio digital robust blind watermark embedding method based on dual-tree complex wavelet transform |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110163787A true CN110163787A (en) | 2019-08-23 |
| CN110163787B CN110163787B (en) | 2023-02-28 |
Family
ID=67640006
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910343271.XA Active CN110163787B (en) | 2019-04-26 | 2019-04-26 | Audio digital robust blind watermark embedding method based on dual-tree complex wavelet transform |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110163787B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112927700A (en) * | 2021-02-06 | 2021-06-08 | 兰州理工大学 | Blind audio watermark embedding and extracting method and system |
| CN113470666A (en) * | 2021-06-21 | 2021-10-01 | 南京信息工程大学 | Reversible robust medical audio method based on two-stage embedding |
| CN114168908A (en) * | 2020-09-11 | 2022-03-11 | 四川大学 | Copyright protection technology based on audio and video analysis |
| CN115602179A (en) * | 2022-11-28 | 2023-01-13 | 腾讯科技（深圳）有限公司（Cn） | Audio watermark processing method and device, computer equipment and storage medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101039371A (en) * | 2006-03-18 | 2007-09-19 | 辽宁师范大学 | Novel method of digital watermarking for protecting literary property of presswork |
| CN102880997A (en) * | 2011-05-26 | 2013-01-16 | 江苏技术师范学院 | Method for embedding watermark image |
| CN106504757A (en) * | 2016-11-09 | 2017-03-15 | 天津大学 | A kind of adaptive audio blind watermark method based on auditory model |
-
2019
- 2019-04-26 CN CN201910343271.XA patent/CN110163787B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101039371A (en) * | 2006-03-18 | 2007-09-19 | 辽宁师范大学 | Novel method of digital watermarking for protecting literary property of presswork |
| CN102880997A (en) * | 2011-05-26 | 2013-01-16 | 江苏技术师范学院 | Method for embedding watermark image |
| CN106504757A (en) * | 2016-11-09 | 2017-03-15 | 天津大学 | A kind of adaptive audio blind watermark method based on auditory model |
Non-Patent Citations (1)
| Title |
|---|
| 张敏情等: "结合视觉感知特性和湿纸编码的视频隐写算法", 《计算机应用研究》 * |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114168908A (en) * | 2020-09-11 | 2022-03-11 | 四川大学 | Copyright protection technology based on audio and video analysis |
| CN112927700A (en) * | 2021-02-06 | 2021-06-08 | 兰州理工大学 | Blind audio watermark embedding and extracting method and system |
| CN112927700B (en) * | 2021-02-06 | 2024-03-19 | 兰州理工大学 | Blind audio watermark embedding and extracting method and system |
| CN113470666A (en) * | 2021-06-21 | 2021-10-01 | 南京信息工程大学 | Reversible robust medical audio method based on two-stage embedding |
| CN113470666B (en) * | 2021-06-21 | 2023-05-16 | 南京信息工程大学 | Reversible robust medical audio method based on two-stage embedding |
| CN115602179A (en) * | 2022-11-28 | 2023-01-13 | 腾讯科技（深圳）有限公司（Cn） | Audio watermark processing method and device, computer equipment and storage medium |
| CN115602179B (en) * | 2022-11-28 | 2023-03-24 | 腾讯科技（深圳）有限公司 | Audio watermark processing method and device, computer equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110163787B (en) | 2023-02-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110163787A (en) | Digital audio Robust Blind Watermarking Scheme embedding grammar based on dual-tree complex wavelet transform | |
| Wu et al. | Efficiently self-synchronized audio watermarking for assured audio data transmission | |
| KR100595202B1 (en) | Apparatus of inserting/detecting watermark in Digital Audio and Method of the same | |
| Dhar et al. | A new audio watermarking system using discrete fourier transform for copyright protection | |
| WO2007148883A1 (en) | Apparatus and method for inserting/extracting capturing-resistant audio watermark based on discrete wavelet transform, audio rights protection system using the same | |
| CN105632506A (en) | Robust digital audio watermark embedding and detection method based on polar harmonic transform | |
| Dhar et al. | Digital watermarking scheme based on fast Fourier transformation for audio copyright protection | |
| KR20030073369A (en) | A Real-time Blind Video Watermarking Method using Quantization | |
| CN1848829B (en) | Method for automatic synchronizing of audio-frequency watermark | |
| CN102237093B (en) | Echo hiding method based on forward and backward echo kernels | |
| Martin et al. | Information retrieval in digital watermarking | |
| Huang et al. | A Fast and Low-Distortion Capacity Adaptive Synchronized Acoustic-to-Acoustic Steganography Scheme | |
| US7886151B2 (en) | Temporal synchronization of video and audio signals | |
| CN104217725A (en) | Audio watermarking method based on multi-echo core | |
| JP5554658B2 (en) | Audio digital watermark embedding apparatus and program | |
| CN102123327A (en) | Method for embedding and extracting digital watermark on basis of streaming media noncritical frame | |
| Fallahpour et al. | DWT-based high capacity audio watermarking | |
| Naik et al. | Joint Encryption and Compression scheme for a multimodal telebiometric system | |
| Dhavale et al. | Walsh Hadamard transform based robust blind watermarking for digital audio copyright protection | |
| Dieu et al. | An improved technique for hiding data in audio | |
| CN108877819A (en) | A kind of voice content evidence collecting method based on coefficient correlation | |
| CN112290975A (en) | Noise estimation receiving method and device for audio information hiding system | |
| Ko et al. | Robust audio watermarking in wavelet domain using pseudorandom sequences | |
| Huang et al. | Design and implementation of synchronized audio-to-audio steganography scheme | |
| Xiong et al. | Optimized SS-based Audio Watermarking System Using Partial Embedding Method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| CB02 | Change of applicant information |
Address after: 210000 Jiangdong North Road, Gulou District, Nanjing, Jiangsu 101 Applicant after: Jiangsu Watermark Technology Co.,Ltd. Address before: 210000 Jiangdong North Road, Gulou District, Nanjing, Jiangsu 101 Applicant before: JIANGSU XINSHI CLOUD SAFETY TECHNOLOGY Co.,Ltd. |
|
| CB02 | Change of applicant information | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |