CN102664013A - Audio digital watermark method of discrete cosine transform domain based on energy selection - Google Patents
Audio digital watermark method of discrete cosine transform domain based on energy selection Download PDFInfo
- Publication number
- CN102664013A CN102664013A CN2012101138448A CN201210113844A CN102664013A CN 102664013 A CN102664013 A CN 102664013A CN 2012101138448 A CN2012101138448 A CN 2012101138448A CN 201210113844 A CN201210113844 A CN 201210113844A CN 102664013 A CN102664013 A CN 102664013A
- Authority
- CN
- China
- Prior art keywords
- discrete cosine
- cosine transform
- watermark
- energy
- segmentation
- 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.)
- Pending
Links
Images
Abstract
The invention discloses an audio digital watermark method of a discrete cosine transform domain based on energy selection, and the method belongs to the technical field of information security. The method comprises the steps of embedding watermarks into audio carrier signals and extracting the watermarks from audio signals into which the watermarks are embedded; methods of time domain sequence segmentation and energy selection based on time domain segmentation are adopted at both watermark embedding and watermark exacting steps; and optimal image watermark embed points are sought in a self-adaptive way through the characteristics of invariable discrete cosine transform energy and energy compression and the audio digital watermark method that the existing image watermark are embedded into low-frequency discrete cosine transform coefficients. Compared with an audio watermark technology without energy selection, the method not only has good concealment, but also has the characteristics of high attack resistance capability to common signal processing and good robustness.
Description
Technical field
The present invention relates to the method that a kind of watermark embeds and extracts, particularly a kind of digital watermark method for audio of the discrete cosine transform domain of selecting based on energy belongs to field of information security technology.
Background technology
Digital watermarking is an important branch in Information hiding (Information Hiding) the technical research field, and it plays an important role in the copyright protection of multimedia messages.
Digital audio frequency watermark is authenticity and the proprietorial a kind of technology that shows sound signal in the sound signal through watermark is embedded into, and it comprises the embedding of watermark and extracts two parts.
Audio frequency watermark should satisfy following requirement: (1) disguise: after sound signal added watermark, added digital watermarking can not influence the quality of former sound signal; (2) robustness: the watermark data of embedding can not be removed or eliminate with attacking through signal processing operations commonly used by unwarranted user; (3) security: security means that watermark can only be measured by the manual inspection that is authorized to.(Pranab?Kumar?Dhar,?Isao?Echizen.?Robust?FFT?Based?Watermarking?Scheme?for?Copyright?Protection?of?Digital?Audio?Data.?2011?Seventh?International?Conference?on?Intelligent?Information?Hiding?and?Multimedia?Signal?Processing,2011:181-184.)
In recent years, there have been a lot of watermarking algorithms to be suggested and have been used for producing strong and hidden audio frequency watermark.Certain methods is embedded in watermark in the time domain of sound signal.Other digital watermark has adopted the method for conversion to come embed watermark, and for example: DFT (DFT), discrete cosine transform (DCT) or discrete wavelet change (DWT).(1, P Bassia; I Pitas; N Nikolaidis. Robust Audio Watermarking in the Time domain. IEEE Transaction on Multimedia; June.2011:vol.3; No.2; Pp.232-241.2, L Xie; J Zhang, H He. Robust Audio Watermarking Scheme Based on Nonuniform Discrete Fourier Transform. in Proceedings of IEEE International Conference on Engineering of Intelligent System, 2006:1-5.3, G Zeng; Z Qiu. Audio Watermarking in DCT:Embedding Strategy and Algorithm. in Proceedings of 9th International Conference on Signal Processing (ICSP'09); 2008:2193-2196.4, M Pooyan, A Delforouzi. Adaptive and Robust Audio watermarking in Wavelet Domain. in Proceedings of International Conference on International Information Hiding and Multimedia Signal Processing (IIH-MSP 2007), 2007:vol. 2; Pp.287-290.) and the discrete cosine transform watermarking algorithm because the less characteristic of its calculated amount, make the scholar more for the research in this field.
The analytical form of discrete cosine transform (DCT) is following:
(2)
In the formula; F is a DCT domain coefficient sequence; F is a time domain sequences,
.
No matter pass through what method realizes watermark on the DCT coefficient embedding, its final influence can think on the DCT coefficient, to be superimposed with a noise signal.So i DCT coefficient behind the embed watermark can be expressed as:
Time domain sequences behind the corresponding embed watermark can be expressed as:
In the formula,
(6)
According to the different values of k in the formula (6), formula (6) can be expressed as:
Summary of the invention
The object of the present invention is to provide a kind of digital watermark method for audio that combines sound signal time domain segmentation energy and audio signal energies to concentrate on the characteristics of low frequency discrete cosine transform coefficient; Obtain a kind of water mark method of science more, make that the embedding point of watermark is to confirm according to the next of size adaptation of sound signal segmentation energy.
The technical solution that realizes the object of the invention is: a kind of digital watermark method for audio of the discrete cosine transform domain of selecting based on energy, and step is following:
The first step; Calculate time domain fragment sequence energy and ordering; Sampling original audio time domain sequences is later carried out segmentation; Original audio time domain sequences after the segmentation is calculated every section energy respectively, each segmentation is sorted, and set up the energy index and come the position number of segmentation before ordering behind the record ordering by the size of segmentation energy;
Second step; Energy is selected; Size according to watermark picture matrix behind the dimensionality reduction; From the discrete cosine transform coefficient array of whole time domain segmentation, select N the bigger corresponding discrete cosine transform coefficient of segmentation of energy to embed RP as the watermark picture according to the energy index, the size of N is by the size decision of one dimension watermark picture array;
In the 3rd step, the embedding of watermark picture according to the numerical value of the every row of matrix of one dimension watermark picture, according to the rule of certain embedding, is revised the numerical value through corresponding low frequency discrete cosine transform coefficient after the discrete cosine transform, reaches the purpose of embed watermark;
The 4th step; The extraction of watermark picture; The sound signal of embed watermark picture and the time domain sequences of original audio signal are carried out segmentation respectively and carried out discrete cosine transform; Utilize the energy index of the original audio time domain sequences after the segmentation, from both discrete cosine transform coefficient arrays, select corresponding low frequency coefficient to adjudicate, thereby realize the extraction of watermark picture by the size of segmentation energy.
The present invention compared with prior art; Its remarkable advantage: according to human auditory system's masking characteristics; Utilize the characteristic of discrete cosine transform constant in energy, select the adaptive best watermark picture insertion point of in the discrete cosine transform domain low frequency coefficient, seeking through the time domain sequences energy.
Below in conjunction with accompanying drawing the present invention is described in further detail.
Description of drawings
Fig. 1 is that the inventive method watermark embeds process flow diagram
Fig. 2 is the inventive method watermark extracting process flow diagram
Fig. 3 is the original audio signal oscillogram
Fig. 4 is the sound signal oscillogram behind the embed watermark picture
Fig. 5 is the result who the present invention is carried out robustness test gained.
Embodiment
In conjunction with Fig. 1, Fig. 2, the present invention is based on the digital watermark method for audio of the discrete cosine transform domain of energy selection, step is following:
The first step; Calculate time domain fragment sequence energy and ordering; Sampling original audio time domain sequences is later carried out segmentation, the original audio time domain sequences after the segmentation is calculated every section energy respectively, each segmentation is sorted by the size of segmentation energy; And set up the energy index and come the position number of segmentation before ordering behind the record ordering, detailed process is following:
(1) original audio signal is carried out segmentation, read in original audio signal
, the size of L is by time length of field and the SF decision of sound signal,
Be time-domain audio signal each numerical value behind over-sampling, A is carried out segmentation, every section comprises
L1 sampled point,
, sound signal is divided into
L2 sections,
, obtain two-dimentional audio frequency array
,
Be through dividing in the two-dimentional audio frequency array in back the numerical value of each;
(2) calculating energy,
calculates the energy of each time domain segmentation according to formula;
(3) set up the energy index, the segmentation of time domain sequences is sorted by energy from big to small, set up the position number of each segmentation before energy ordering after energy index index comes record ordering;
Second step; Energy is selected; According to the size of watermark picture matrix behind the dimensionality reduction, from the discrete cosine transform coefficient array of whole time domain segmentation, select N the bigger corresponding discrete cosine transform coefficient of segmentation of energy to embed RP according to the energy index as the watermark picture, detailed process is following:
(1) reads in the binary watermarking picture
,
Be each numerical value in the two-dimentional watermark picture array,
M1 is the line number of two-dimentional watermark picture array,
M2 is the columns of two-dimentional watermark picture array, carries out watermark for ease and embeds, and two-dimension picture array dimensionality reduction to one dimension, is obtained
,
Be each numerical value the watermark picture array after two dimension is reduced to one dimension;
(2) discrete cosine transform is carried out in each segmentation; Obtain discrete cosine transform coefficient array
,
carry out after the discrete cosine transform numerical value of each in the discrete cosine transform coefficient array for each time-domain audio segmentation;
(3) from the discrete cosine transform coefficient array, select the embedding RP of N bigger low frequency discrete cosine transform coefficient of energy as the watermark picture,
according to energy index index;
In the 3rd step, the embedding of watermark picture is according to the numerical value of the every row of matrix of one dimension watermark picture; Rule according to certain embedding; Revise the numerical value through corresponding low frequency discrete cosine transform conversion coefficient after the discrete cosine transform, reach the purpose of embed watermark, detailed process is following:
(1) with one dimension watermark picture matrix numerical value binaryzation;
is to improve the robustness of embed watermark;
(2) carry out watermark according to certain rule and embed,
,
Be
hThe low frequency DCT coefficient of individual segmentation,
Be the value of embed watermark,
It is scale factor;
The 4th step; The extraction of watermark picture; The sound signal of embed watermark picture and the time domain sequences of original audio signal are carried out segmentation respectively and carried out discrete cosine transform, utilize the energy index of the original audio time domain sequences after the segmentation, from both discrete cosine transform coefficient arrays, select corresponding low frequency coefficient to adjudicate by the size of segmentation energy; Thereby realize the extraction of watermark picture, detailed process is following:
(1) sound signal behind the embed watermark picture is carried out segmentation; Read in the sound signal
behind the embed watermark;
is the numerical value of each afterwards that the time-domain audio signal behind the embed watermark picture is sampled; D is carried out segmentation; Every section comprises L1 sampled point;
; Sound signal is divided into the L2 section;
; Obtain two-dimentional audio frequency array
,
for
being divided in the two-dimentional audio frequency array in back the numerical value of each;
(2) discrete cosine transform is carried out in each time domain segmentation of
; Obtain discrete cosine transform coefficient array
,
carry out after the discrete cosine transform numerical value of each in the discrete cosine transform coefficient array for the time-domain audio segmentation behind each embed watermark picture;
(3) according to the energy index index of original audio signal; From C and E, select corresponding DCT low frequency coefficient to carry out the size judgement respectively; If the numerical value of discrete cosine transform low frequency coefficient among the E>numerical value of discrete cosine transform low frequency coefficient among the C; Judge that then this watermark picture numerical matrix numerical value is 1, if among the E numerical value of discrete cosine transform low frequency coefficient < numerical value of discrete cosine transform low frequency coefficient among the C judges that then this watermark picture numerical matrix numerical value is 0;
(4) one dimension watermark picture matrix is upgraded to two-dimensional matrix, realizes the extraction of embed watermark picture.
In order to verify the effect of the inventive method, carry out following experiment respectively:
(1) hides property testing: detect the hiding property of the inventive method for the watermark picture that embeds.It is 22.05kHz that original audio signal adopts SF as shown in Figure 3, and sample size is 16, and position speed is 352kbps, and duration is the wav audio file of 5.0363s, and sampling number is 111051.When getting 0.1, the audio file oscillogram behind the embed watermark image is as shown in Figure 4 as
.
(2) robustness test: the damage situations that detects the inventive method and the classic method of not selecting through energy embed watermark when standing various common signal Processing attack.The institute's white noise that under fire is respectively attack, filtering attack and weight attack.As shown in Figure 5, A is an original watermark, the watermark extracting effect of B1 the inventive method when not adding any attack, the watermark extracting effect of B2 classic method when not adding any attack; The watermark extracting effect of the inventive method when C1 is the additive white Gaussian noise attack, the watermark extracting effect of classic method when C2 is the additive white Gaussian noise attack; The watermark extracting effect of the inventive method when D1 attacks for the low pass Butterworth filter, the watermark extracting effect of classic method when D2 attacks for the low pass Butterworth filter; The watermark extracting effect of the inventive method when E1 is the attack of 8bit weight, the watermark extracting effect of classic method when E2 is the attack of 8bit weight.
Claims (5)
1. the digital watermark method for audio of a discrete cosine transform domain of selecting based on energy is characterized in that comprising following steps:
The first step; Calculate time domain fragment sequence energy and ordering; Sampling original audio time domain sequences is later carried out segmentation; Original audio time domain sequences after the segmentation is calculated every section energy respectively, each segmentation is sorted, and set up the energy index and come the position number of segmentation before ordering behind the record ordering by the size of segmentation energy;
Second step; Energy is selected; Size according to watermark picture matrix behind the dimensionality reduction; From the discrete cosine transform coefficient array of whole time domain segmentation, select N the bigger corresponding discrete cosine transform coefficient of segmentation of energy to embed RP as the watermark picture according to the energy index, the size of N is by the size decision of one dimension watermark picture array;
In the 3rd step, the embedding of watermark picture according to the numerical value of the every row of matrix of one dimension watermark picture, according to the rule of certain embedding, is revised the numerical value through corresponding low frequency discrete cosine transform coefficient after the discrete cosine transform, reaches the purpose of embed watermark;
The 4th step; The extraction of watermark picture; The sound signal of embed watermark picture and the time domain sequences of original audio signal are carried out segmentation respectively and carried out discrete cosine transform; Utilize the energy index of the original audio time domain sequences after the segmentation, from both discrete cosine transform coefficient arrays, select corresponding low frequency coefficient to adjudicate, thereby realize the extraction of watermark picture by the size of segmentation energy.
2. the digital watermark method for audio of the discrete cosine transform domain of selecting based on energy according to claim 1 is characterized in that calculating time domain fragment sequence energy and ordering, and detailed process is following:
The first step is carried out segmentation to original audio signal, reads in original audio signal
,
LSize by time length of field and the SF decision of sound signal,
Be time-domain audio signal each numerical value behind over-sampling, A is carried out segmentation, every section comprises
L1 sampled point,
, sound signal is divided into
L2 sections,
, obtain two-dimentional audio frequency array
,
Be right
The numerical value of each in the two-dimentional audio frequency array after process is divided;
Second step; Calculating energy,
calculates the energy of each time domain segmentation according to formula;
The 3rd step, set up the energy index, the segmentation of time domain sequences is sorted by energy from big to small, set up the position of each segmentation before energy ordering after energy index index comes record ordering.
3. the digital watermark method for audio of the discrete cosine transform domain of selecting based on energy according to claim 1 is characterized in that energy selects, and detailed process is following:
The first step is read in the binary watermarking picture
,
Be each numerical value in the two-dimentional watermark picture array,
M1 is the line number of two-dimentional watermark picture array,
M2 is the columns of two-dimentional watermark picture array, carries out watermark for ease and embeds, and two-dimension picture array dimensionality reduction to one dimension, is obtained
,
Be each numerical value the watermark picture array after two dimension is reduced to one dimension;
Second step; Discrete cosine transform is carried out in each segmentation; Obtain discrete cosine transform coefficient array
,
carry out after the discrete cosine transform numerical value of each in the discrete cosine transform coefficient array for each time-domain audio segmentation;
In the 3rd step, from the discrete cosine transform coefficient array, select energy bigger according to energy index index
NIndividual low frequency discrete cosine transform coefficient is as the embedding RP of watermark picture,
4. the digital watermark method for audio of the discrete cosine transform domain of selecting based on energy according to claim 1 is characterized in that the embedding of watermark picture, and detailed process is following:
The first step; With one dimension watermark picture matrix numerical value binaryzation;
is to improve the robustness of embed watermark;
Second step, carry out watermark according to certain rule and embed,
,
Be
hThe low frequency discrete cosine transform coefficient of individual segmentation,
Be the value of embed watermark,
It is scale factor.
5. the digital watermark method for audio of the discrete cosine transform domain of selecting based on energy according to claim 1 is characterized in that the extraction of watermark picture, and detailed process is following:
The first step is carried out segmentation to the sound signal behind the embed watermark picture, reads in the sound signal behind the embed watermark
,
For the numerical value of each afterwards that the time-domain audio signal behind the embed watermark picture is sampled, D is carried out segmentation, every section comprises
L1 sampled point,
L1=16, sound signal is divided into
L2 sections,
L2=
L/
L1, obtain two-dimentional audio frequency array
,
For right
The numerical value of each in the two-dimentional audio frequency array after dividing;
Second step; Discrete cosine transform is carried out in each time domain segmentation to
; Obtain discrete cosine transform coefficient array
,
carry out after the discrete cosine transform numerical value of each in the discrete cosine transform coefficient array for the time-domain audio segmentation behind each embed watermark picture;
The 3rd step; Energy index index according to original audio signal; From C and E, select corresponding discrete cosine transform low frequency coefficient to carry out size judgement respectively, as if the numerical value of discrete cosine transform low frequency coefficient among the E>numerical value of discrete cosine transform low frequency coefficient among the C, judge that then this watermark picture numerical matrix numerical value is 1; If < numerical value of discrete cosine transform low frequency coefficient among the C judges that then this watermark picture numerical matrix numerical value is 0 to the numerical value of discrete cosine transform low frequency coefficient among the E;
The 4th step was upgraded to two-dimensional matrix with one dimension watermark picture matrix, realized the extraction of embed watermark picture.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2012101138448A CN102664013A (en) | 2012-04-18 | 2012-04-18 | Audio digital watermark method of discrete cosine transform domain based on energy selection |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2012101138448A CN102664013A (en) | 2012-04-18 | 2012-04-18 | Audio digital watermark method of discrete cosine transform domain based on energy selection |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN102664013A true CN102664013A (en) | 2012-09-12 |
Family
ID=46773479
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2012101138448A Pending CN102664013A (en) | 2012-04-18 | 2012-04-18 | Audio digital watermark method of discrete cosine transform domain based on energy selection |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102664013A (en) |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104810022A (en) * | 2015-05-11 | 2015-07-29 | 东北师范大学 | Time-domain digital audio watermarking method based on audio breakpoint |
| CN105787444A (en) * | 2016-02-24 | 2016-07-20 | 北方工业大学 | Signal denoising method based on V system |
| CN106409302A (en) * | 2016-06-22 | 2017-02-15 | 华中师范大学 | Audio frequency watermark method and system based on embedding area selection |
| CN106898358A (en) * | 2017-03-07 | 2017-06-27 | 武汉大学 | From the robust digital audio watermark algorithm of time frequency analysis angle |
| CN109829846A (en) * | 2019-01-29 | 2019-05-31 | 鲁东大学 | A kind of digital picture blind watermark method based on two-dimension discrete cosine transform |
| CN111292756A (en) * | 2020-01-19 | 2020-06-16 | 成都嗨翻屋科技有限公司 | Compression-resistant audio silent watermark embedding and extracting method and system |
| CN111625786A (en) * | 2020-05-07 | 2020-09-04 | 清华四川能源互联网研究院 | Time sequence data watermarking algorithm based on discrete cosine transform |
| CN112735426A (en) * | 2020-12-24 | 2021-04-30 | 深圳市声扬科技有限公司 | Voice verification method and system, computer device and storage medium |
| CN114168908A (en) * | 2020-09-11 | 2022-03-11 | 四川大学 | Copyright protection technology based on audio and video analysis |
| CN114420111A (en) * | 2022-03-31 | 2022-04-29 | 成都启英泰伦科技有限公司 | One-dimensional hypothesis-based speech vector distance calculation method |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1529246A (en) * | 2003-09-28 | 2004-09-15 | 王向阳 | Digital audio-frequency water-print inlaying and detecting method based on auditory characteristic and integer lift ripple |
| WO2005002200A2 (en) * | 2003-06-13 | 2005-01-06 | Nielsen Media Research, Inc. | Methods and apparatus for embedding watermarks |
| CN101345054A (en) * | 2008-08-25 | 2009-01-14 | 苏州大学 | Digital watermark production and recognition method used for audio document |
| CN101345052A (en) * | 2008-08-20 | 2009-01-14 | 回静 | Imbedding method for invisible digital watermark |
| CN102238388A (en) * | 2011-06-29 | 2011-11-09 | 西安电子科技大学 | Self-adaptive robust video watermarking method based on AVS (Audio Video Standard) |
-
2012
- 2012-04-18 CN CN2012101138448A patent/CN102664013A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2005002200A2 (en) * | 2003-06-13 | 2005-01-06 | Nielsen Media Research, Inc. | Methods and apparatus for embedding watermarks |
| CN1529246A (en) * | 2003-09-28 | 2004-09-15 | 王向阳 | Digital audio-frequency water-print inlaying and detecting method based on auditory characteristic and integer lift ripple |
| CN101345052A (en) * | 2008-08-20 | 2009-01-14 | 回静 | Imbedding method for invisible digital watermark |
| CN101345054A (en) * | 2008-08-25 | 2009-01-14 | 苏州大学 | Digital watermark production and recognition method used for audio document |
| CN102238388A (en) * | 2011-06-29 | 2011-11-09 | 西安电子科技大学 | Self-adaptive robust video watermarking method based on AVS (Audio Video Standard) |
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104810022B (en) * | 2015-05-11 | 2018-06-15 | 东北师范大学 | A kind of time domain digital audio water mark method based on audio breakpoint |
| CN104810022A (en) * | 2015-05-11 | 2015-07-29 | 东北师范大学 | Time-domain digital audio watermarking method based on audio breakpoint |
| CN105787444A (en) * | 2016-02-24 | 2016-07-20 | 北方工业大学 | Signal denoising method based on V system |
| CN106409302A (en) * | 2016-06-22 | 2017-02-15 | 华中师范大学 | Audio frequency watermark method and system based on embedding area selection |
| CN106409302B (en) * | 2016-06-22 | 2019-07-09 | 华中师范大学 | Audio-frequency water mark method and system based on insertion regional choice |
| CN106898358A (en) * | 2017-03-07 | 2017-06-27 | 武汉大学 | From the robust digital audio watermark algorithm of time frequency analysis angle |
| CN106898358B (en) * | 2017-03-07 | 2020-01-24 | 武汉大学 | Robust digital audio watermarking algorithm from time-frequency analysis angle |
| CN109829846B (en) * | 2019-01-29 | 2023-07-07 | 鲁东大学 | Digital image blind watermarking method based on two-dimensional discrete cosine transform |
| CN109829846A (en) * | 2019-01-29 | 2019-05-31 | 鲁东大学 | A kind of digital picture blind watermark method based on two-dimension discrete cosine transform |
| CN111292756A (en) * | 2020-01-19 | 2020-06-16 | 成都嗨翻屋科技有限公司 | Compression-resistant audio silent watermark embedding and extracting method and system |
| CN111292756B (en) * | 2020-01-19 | 2023-05-26 | 成都潜在人工智能科技有限公司 | Compression-resistant audio silent watermark embedding and extracting method and system |
| CN111625786A (en) * | 2020-05-07 | 2020-09-04 | 清华四川能源互联网研究院 | Time sequence data watermarking algorithm based on discrete cosine transform |
| CN114168908A (en) * | 2020-09-11 | 2022-03-11 | 四川大学 | Copyright protection technology based on audio and video analysis |
| CN112735426A (en) * | 2020-12-24 | 2021-04-30 | 深圳市声扬科技有限公司 | Voice verification method and system, computer device and storage medium |
| CN114420111A (en) * | 2022-03-31 | 2022-04-29 | 成都启英泰伦科技有限公司 | One-dimensional hypothesis-based speech vector distance calculation method |
| CN114420111B (en) * | 2022-03-31 | 2022-06-17 | 成都启英泰伦科技有限公司 | One-dimensional hypothesis-based speech vector distance calculation method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102664013A (en) | Audio digital watermark method of discrete cosine transform domain based on energy selection | |
| CN102917227B (en) | Compressive sensing-based adaptive video information hiding method | |
| Dhar et al. | A new audio watermarking system using discrete fourier transform for copyright protection | |
| CN106898358B (en) | Robust digital audio watermarking algorithm from time-frequency analysis angle | |
| CN101908201A (en) | Robust digital image adaptive zero-watermarking method | |
| Dhar et al. | Digital watermarking scheme based on fast Fourier transformation for audio copyright protection | |
| Pun | A novel DFT-based digital watermarking system for images | |
| CN111028850A (en) | Audio watermark embedding method and audio watermark extracting method | |
| CN106780281A (en) | Digital image watermarking method based on Cauchy's statistical modeling | |
| CN104978706B (en) | Watermark embedding and extracting method, device and system based on Fourier descriptor | |
| CN103886864A (en) | Method for detecting quantitatively-embedded secret information for DWT domain fidelity compensation | |
| Li et al. | Robust audio watermarking scheme via QIM of correlation coefficients using LWT and QR decomposition | |
| Cetinel et al. | Robust chaotic digital image watermarking scheme based on RDWT and SVD | |
| CN103079066A (en) | Image watermark embedding and extracting method combining ridgelet transform and DCT transform | |
| CN102496371B (en) | Digital watermark method for audio carrier | |
| CN107910010B (en) | Digital watermark detection method based on multi-parameter Weibull statistical modeling | |
| Bhat K et al. | Audio watermarking based on quantization in wavelet domain | |
| CN103871015B (en) | Watermark method aiming at computer graphic image | |
| CN108877819B (en) | Voice content evidence obtaining method based on coefficient autocorrelation | |
| Mothi et al. | A wavelet packet and fuzzy based digital image watermarking | |
| Chen et al. | A multipurpose audio watermarking scheme for copyright protection and content authentication | |
| Nafornita et al. | A wavelet-based digital watermarking using subband adaptive thresholding for still images | |
| Ma et al. | A robust watermarking scheme based on dual quantization of wavelet significant difference | |
| CN109712059B (en) | Digital watermark detection method based on multi-scale joint statistical modeling | |
| CN110047495A (en) | Large capacity Audio Watermarking Algorithm based on 2 grades of singular value decompositions |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20120912 |
|
| WD01 | Invention patent application deemed withdrawn after publication |