CN110010142B - Large-capacity audio information hiding method - Google Patents
Large-capacity audio information hiding method Download PDFInfo
- Publication number
- CN110010142B CN110010142B CN201910243836.7A CN201910243836A CN110010142B CN 110010142 B CN110010142 B CN 110010142B CN 201910243836 A CN201910243836 A CN 201910243836A CN 110010142 B CN110010142 B CN 110010142B
- Authority
- CN
- China
- Prior art keywords
- wavelet coefficient
- audio
- hidden information
- carrier
- wavelet
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired - Fee Related
Links
- 238000000034 method Methods 0.000 title claims abstract description 41
- 230000008569 process Effects 0.000 claims abstract description 12
- 238000012937 correction Methods 0.000 claims description 25
- 229910002056 binary alloy Inorganic materials 0.000 claims description 9
- 238000006243 chemical reaction Methods 0.000 claims description 9
- 238000002156 mixing Methods 0.000 claims description 9
- 230000003044 adaptive effect Effects 0.000 claims description 8
- 238000004364 calculation method Methods 0.000 claims description 8
- 238000012545 processing Methods 0.000 claims description 8
- 238000012216 screening Methods 0.000 claims description 6
- 238000000605 extraction Methods 0.000 claims description 5
- 238000010606 normalization Methods 0.000 claims description 2
- 238000005516 engineering process Methods 0.000 description 9
- 230000005236 sound signal Effects 0.000 description 8
- 230000001105 regulatory effect Effects 0.000 description 4
- 230000005540 biological transmission Effects 0.000 description 3
- 238000000354 decomposition reaction Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 238000006467 substitution reaction Methods 0.000 description 3
- 230000009466 transformation Effects 0.000 description 3
- 238000004891 communication Methods 0.000 description 2
- 210000005069 ears Anatomy 0.000 description 2
- 239000000284 extract Substances 0.000 description 2
- 238000001914 filtration Methods 0.000 description 2
- 230000000873 masking effect Effects 0.000 description 2
- 230000008447 perception Effects 0.000 description 2
- 230000001737 promoting effect Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 238000010187 selection method Methods 0.000 description 2
- 239000000126 substance Substances 0.000 description 2
- 230000000007 visual effect Effects 0.000 description 2
- 230000004075 alteration Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 239000000969 carrier Substances 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L19/00—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis
- G10L19/018—Audio watermarking, i.e. embedding inaudible data in the audio signal
Landscapes
- Engineering & Computer Science (AREA)
- Computational Linguistics (AREA)
- Signal Processing (AREA)
- Health & Medical Sciences (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Human Computer Interaction (AREA)
- Physics & Mathematics (AREA)
- Acoustics & Sound (AREA)
- Multimedia (AREA)
- Signal Processing For Digital Recording And Reproducing (AREA)
- Compression, Expansion, Code Conversion, And Decoders (AREA)
Abstract
The invention provides a method for hiding high-capacity audio information, which utilizes a self-adaptive threshold value based on average energy to respectively screen wavelet coefficients of carrier audio and hidden information, so that the portable audio has better transparency; according to the characteristics of actual concealed voice, the self-adaptive capacity selection is realized by utilizing the characteristics of wavelet coefficients, and the efficiency of embedding concealed information into carrier audio is improved; in the process of embedding the hidden information, the hidden information is embedded in an indirect mode, the hidden information is converted into an embedding factor and then hidden in the carrier audio, and then a plurality of low bits of the binary number of the carrier audio are replaced by the binary number of the embedding factor by utilizing the thought of a least important replacement method, so that the secret-carrying information is obtained. The invention has good transparency and robustness, can greatly improve the efficiency of embedding the carrier audio into the hidden information, and has relative capacity of 100 percent or more. The method is simple, high in execution efficiency and suitable for hiding high-capacity audio information.
Description
Technical Field
The invention belongs to the technical field of audio information hiding, and particularly relates to a large-capacity audio information hiding method based on wavelet transformation, which has good transparency and robustness.
Background
Compared with the traditional cryptology technology, the information hiding technology focuses more on the existence of the hidden information. The method and the device embed the secret information into the carrier information by utilizing the redundancy of the carrier information (pictures, voice, video and the like), and the carrier information is hidden to show the external characteristics of the carrier and has limitation on human senses, so that the difference of the carrier information before and after embedding the secret information is not easy to feel, and the purpose of information hiding is achieved. And the carrier with the hidden secret information reaches the receiver through the channel, and the receiver can recover the secret information from the carrier by using the key through the detector.
Audio information hiding techniques face greater challenges than digital images and video watermarking, among others. On one hand, because the human Auditory system HAS (human audio system) is very sensitive to random noise, the amount of information that can be embedded is very limited, and the human Auditory system HAS is much more sensitive than the human Visual system hvs (human Visual system), the imperceptibility of audio information hiding is much more difficult to realize than that of images, which brings great difficulty to the research of audio information hiding; on the other hand, compared with carriers such as video and images, the number of points of audio sampled in the same time interval is much smaller. This means that the amount of information that can be embedded in an audio signal is much smaller than in a carrier such as an image, video, etc. However, in the main human communication methods such as text, picture, and voice, the voice efficiency is higher, so that the audio information hiding technology, which is an information hiding technology using audio information as a carrier, is one of the hot spots of research.
At present, audio information hiding technology is mainly divided into time domain algorithm and transform domain algorithm on the method of hiding information embedding. Classical time domain algorithms include least significant bit substitution, echo concealment, etc. The time domain algorithm is simple to implement but has poor robustness because it only processes the time domain waveform of the audio signal. For example, least significant bit substitution may have good transparency and hidden capacity advantages, but does not stand up to any interference. Even the interference of noise in the transmission channel cannot be resisted, and the low robustness requirement applied to covert communication cannot be met. Although echo hiding improves certain robustness, the hiding capacity is too low and is not blind extraction, which is a big discount for practicality. Therefore, the audio information hiding technology based on the time domain algorithm is greatly limited in practicability. Transform domain algorithms typically first perform some kind of transform (e.g., discrete fourier transform, DFT, discrete cosine transform, discrete wavelet transform, DWT) operation on the carrier signal, and then embed the information by modifying the transform domain coefficients. Because the coefficient obtained by the audio signal in the transform domain can reflect the characteristics of the signal, and the auditory effect and the masking characteristic curve of human ears are combined, the robustness of the algorithm is enhanced, the transparency of the secret audio is greatly improved, but the capacity of the hidden information is usually limited. Therefore, how to increase the capacity of the audio information hiding technology on the basis of ensuring good transparency and robustness is one of the important factors for promoting the practical application of the technology, and becomes a technical problem to be solved in the field.
Disclosure of Invention
In order to solve the above problems, the present invention provides a method for hiding large-capacity audio information.
The technical scheme provided by the invention provides a method for hiding high-capacity audio information, which is characterized in that wavelet coefficients of carrier audio and hidden information are respectively screened by utilizing an adaptive threshold based on average energy, so that the portable audio has better transparency; according to the characteristics of actual concealed voice, the self-adaptive capacity selection is realized by utilizing the characteristics of wavelet coefficients, and the efficiency of embedding concealed information into carrier audio is improved; in the process of embedding the hidden information, the hidden information is embedded in an indirect mode, the hidden information is converted into an embedding factor and then hidden in the carrier audio, and then a plurality of low bits of the binary number of the carrier audio are replaced by the binary number of the embedding factor by utilizing the thought of a least important replacement method, so that the secret-carrying information is obtained.
Moreover, wavelet coefficients of carrier audio and hidden information are respectively screened by utilizing an adaptive threshold based on average energy, which is realized by the following steps,
respectively calculating the average energy of wavelet coefficients of each layer of carrier audio and hidden information;
respectively calculating self-adaptive threshold values of wavelet coefficients of each layer of carrier audio and hidden information according to the average energy and corresponding threshold value correction factors;
comparing the amplitude value of each layer of wavelet coefficient of the carrier audio and the hidden information with the corresponding self-adaptive threshold, if the amplitude value of the wavelet coefficient is larger than the threshold, retaining the amplitude value, otherwise, the wavelet coefficient is 0, thereby realizing the screening of the corresponding wavelet coefficient.
Moreover, the adaptive capacity selection is implemented as follows,
after screening the wavelet coefficients of the carrier audio and the hidden information, obtaining a carrier voice wavelet coefficient X [ k ] after statistical correction and a wavelet coefficient H [ k ] of the hidden information after correction through descending order arrangement and normalization, wherein the number of the carrier voice wavelet coefficient X [ k ] after statistical correction larger than zero is xp, the number of the carrier voice wavelet coefficient X [ k ] after statistical correction smaller than zero is xn, the number of the wavelet coefficient H [ k ] after statistical correction larger than zero is hp, and the number of the carrier voice wavelet coefficient H [ k ] after statistical correction smaller than zero is hn;
when xp + xn is less than hp + hn, gradually increasing the length of the carrier audio, and re-processing the carrier audio until the condition that xp + xn is more than or equal to hp + hn is met;
when xp + xn > > hp + hn, gradually reducing the length of the carrier audio, and repeating the processing related to the carrier audio until the conditions that xp + xn ≧ hp + hn and xp + xn > > hp + hn are no longer satisfied.
Furthermore, the covert information embedding process includes the steps of,
1) let X [ k ] k]Has a length of LX，H[k]Has a length of LH；
Let X [ i ] be the hp item before the carrier voice wavelet coefficient X [ k ] after correction, Hi be the hp item before the hidden information wavelet coefficient Hk after correction, i be the serial number of the hidden information positive wavelet coefficient,
binary system for calculating embedding factor of positive wavelet coefficient
Wherein dec2bin (·) denotes decimal conversion to binary computation;
let X [ j ]]For modified carrier speech wavelet coefficient X k]The next hn item, j is a serial number; h [ j ]1]For the wavelet coefficient H k of the corrected hidden information]Post hn item, j1Is a serial number; j is LX-hn-1、...LX，j1＝LH-hn-1、...LH
Binary system for calculating embedding factor of negative wavelet coefficient
2) Calculate X [ i ]]Binary X ofp[i]And X [ j ]]Binary X ofn[j]；
Mixing Xp[i]Last 5 bit of (A) is replaced by XHp[i]Then binary conversion is carried out to decimal calculation to obtain hp item X 'in front of carrier voice wavelet coefficient containing hidden information'p[i]，
Mixing Xn[j]Last 5 bit of (A) is replaced by XHn[j]Then binary conversion is carried out to decimal calculation to obtain hn item X 'after the carrier voice wavelet coefficient containing the hidden information'n[j]；
3) When descending order arrangement is set, the wavelet coefficients of the carrier voice are spliced end to end and put into an array x [ k ]]In, for x [ k ]]The descending order is arranged to obtain a new array x' [ k ]]X' k is]The front hp wavelet coefficient of (2) is replaced by X'p[i]X' k is]The post-hn wavelet coefficient of (2) is replaced with X'n[j]Corresponding x' k according to the stored position]The position of the middle element is restored to the position of the original wavelet coefficient, and the wavelet coefficient X embedded with the hidden information is obtainedinsert[k]；
4) By using
Mixing Xinsert[k]The value of (a) is restored to the order of magnitude of the original wavelet coefficient value, the coefficient is split according to the length to the result, then the carrying-secret audio x is obtained by inverse wavelet transformationinsert[n]。
Moreover, the extraction of the hidden information corresponds to the hidden information embedding process, comprising the following steps,
1) the received secret-carrying audio yinsert[n]Wavelet transform of 3 layers is carried out to obtain wavelet coefficient y1,H[k]、y2,H[k]、y3,H[k]And y3,L[k]Splicing them end to end into an array y [ k ]]Then use
Normalizing the amplitude to [ - (2)15-1)，(215-1)]The interval is rounded nearby to obtain the wavelet coefficient Y [ k ] of the regular secret-carrying audio]；
2) Corresponding Y [ k ] according to the stored position]The largest hp elements in the array are taken out and arranged in descending order to form a new array Yp[i]And the smallest hn elements are taken out and arranged in descending order to form a new array Yn[j]And the values of these elements are converted into binary, then the last 5 bits of each binary element are extracted and converted into the embedding factor YH of the decimal hidden information positive wavelet coefficientp[i]Embedding factor YH of negative wavelet coefficient of sum hidden informationn[j]；
3) By using
Obtaining the extracted hidden information positive wavelet coefficient Hp[i]By using
Obtaining the negative wavelet coefficient H of the extracted hidden informationn[j]Then form the extracted hidden information wavelet coefficient in descending order
4) According to the corresponding relation of the recorded wavelet coefficients at the original position and the positions arranged in descending order, the H is divided intodesc[k]Converting into extracted hidden information wavelet coefficient Hextra[k]And according to the corresponding length, Hextra[k]Splitting into a first layer high-frequency coefficient h'1,H[k]And a second layer high-frequency coefficient h'2,H[k]And the third layer has a high-frequency coefficient h'3,H[ k ] and third layer Low frequency coefficient h'3,L[k]Then, 3-layer inverse wavelet transform is carried out to obtain hidden information hertra[n]。
The invention has the beneficial effects that: the invention provides a self-adaptive decision threshold value calculation method based on wavelet transformation, which indirectly embeds hidden information by using the least important replacement thought to improve the transparency and robustness of a secret audio, and provides a self-adaptive capacity selection method by using the characteristics of a wavelet coefficient, thereby greatly improving the efficiency of embedding the hidden information into a carrier audio, and the relative capacity reaches 100 percent or more, and the method is a high-capacity audio information hiding method based on wavelet transformation with good transparency and robustness. The method is simple, has high execution efficiency, is suitable for hiding high-capacity audio information, and provides a powerful guarantee for promoting the practicability of the audio information hiding technology.
Drawings
FIG. 1 is a schematic diagram of a system model for audio information hiding according to an embodiment of the present invention;
FIG. 2 is a diagram of the Mallat algorithm of the prior art;
fig. 3 is a block diagram of hidden information embedding according to an embodiment of the present invention.
Detailed Description
The technical scheme of the invention is specifically described in the following with reference to the accompanying drawings and embodiments.
First, the relevant theoretical basis is introduced:
the information hiding emphasizes the existence of the information, namely the hidden information is hidden in the carrier information and is not easy to be found by using the characteristics of the carrier information through some special processing. The voice signal has the redundancy characteristic, the audio information hiding fully utilizes the limitation of auditory perception of human ears, the auditory masking effect is considered, the hidden information is hidden in the carrier audio signal, and an audio information hiding system model adopted by the invention is shown in figure 1.
As shown in fig. 1, the carrier information serves as a carrier for "transporting" covert information, typically digital voice or music signals. The hidden information is all information that can be converted into a binary bit stream, typically text, pictures, voice or video. The implementation process of the audio information hiding system comprises the following steps: first, the hidden information is embedded into the carrier audio by an embedding algorithm, which is usually implemented in the time domain, the frequency domain of the digital signal. The embedded key corresponds to the extracted key. The secret keys are divided into two types, one type is used for enhancing the safety, and after being intercepted by an interceptor, the hidden information is prevented from being decoded into the hidden information; the other is that the key is needed for extracting the hidden information due to the embedding algorithm itself, and the hidden information cannot be extracted without the key. The speech carrying the hidden information, which is not perceptually different from the original carrier audio in terms of auditory perception, is then transmitted over a transmission channel such as the Internet, telephone, radio, etc. Finally, the extraction algorithm extracts the hidden information using the secret key.
The method takes the audio as a carrier, indirectly embeds the hidden information by self-adaptively selecting the decision threshold and utilizing the least important replacement method idea, thereby improving the transparency and the robustness of the secret-carrying audio; and according to the characteristics of the actual concealed voice, the method for selecting the adaptive capacity is provided by utilizing the characteristics of the wavelet coefficient, so that the efficiency of embedding concealed information into the carrier audio is greatly improved, and the relative capacity reaches 100% or more.
1.1Mallat Algorithm
According to the multiresolution theory, a fast algorithm (Mallat algorithm) for wavelet decomposition of one-dimensional discrete signals is shown in FIG. 2, where ↓2in FIG. 2 is downsampled by 2 times, where x [ n ]]The length is N, and N is a time sequence number; g [ n ]]A low pass filter for filtering out a high frequency portion of the input signal and outputting a low frequency portion; h [ n ]]A high-pass filter for filtering out a low-frequency part and outputting a high-frequency part, which is opposite to the low-pass filter; x is the number of1,H[k]Is the first layer high frequency coefficient, x2,H[k]Is the second layer high frequency coefficient, x3,H[k]Is the third layer high frequency coefficient, x3,L[k]Is the third layer low frequency coefficient, k is the number:
1.2 hidden information embedding
For wavelet analysis of audio signals, it is the wavelet coefficients obtained by wavelet transform that most reflect the characteristics of the audio signals. The wavelet coefficients represent the energy of the audio signal, and since the audio signal is mainly concentrated in the low frequency part, the wavelet coefficients tend to have larger values in the low frequency part, so that the hidden information can be hidden in the wavelet coefficients. The process of embedding hidden information is shown in FIG. 3, where the number of X [ k ] greater than zero is xp, the number of X [ k ] less than zero is xn, the number of H [ k ] greater than zero is hp, and the number of H [ k ] less than zero is hn.
In the hidden information embedding method, in order to improve the transparency and the robustness of the secret-carrying audio, the invention improves the prior art and provides a self-adaptive selection judgment threshold value; in order to improve the efficiency of embedding hidden information into carrier audio, the invention provides a self-adaptive capacity selection method.
The embodiment provides a method for hiding high-capacity audio information, which comprises the following steps:
step 1, data x [ n ] of one channel of carrier audio and hidden information hd [ n ] are collected, wherein n is a serial number.
Step 2, for carrier audio x [ n ]]And hidden information hd n]Wavelet coefficient x obtained by performing 3-layer discrete wavelet decomposition1,H[k]、x2,H[k]、x3,H[k]And x3,L[k]And k is a serial number, and the lengths of the wavelet coefficients of each group are stored in variables Lx1, Lx2, Lx3 and Lx4 respectively in sequence. For hidden information hd [ n ]]Wavelet coefficient h obtained by 3-layer discrete wavelet decomposition1,H[k]、h2,H[k]、h3,H[k]And h3,L[k]The lengths of the wavelet coefficients of each group are stored in the variables Lh1, Lh2, Lh3 and Lh4, respectively, in order.
Step 3, utilizing the self-adaptive threshold value based on average energy to respectively screen the wavelet coefficients of the carrier audio and the hidden information, so that the portable audio has better transparency, and the specific implementation steps are as follows,
step 3.1, use
Calculating the average energy of wavelet coefficient of each layer of carrier audio frequency and hidden information respectively, E [ k ]]And M is the number of corresponding wavelet coefficients. Such as: calculating the average energy of the high-frequency coefficient of the first layer of the carrier audio frequency, and then E [ k ]]＝x1,H[k]And M is the number of the high-frequency coefficients of the first layer of the carrier audio.
Step 3.2, using TH ═ α × E to respectively calculate adaptive thresholds of wavelet coefficients of each layer of carrier audio and hidden information, where α is a threshold correction factor corresponding to the wavelet coefficient, and a suggested preferred value obtained through experiments is: threshold correction factor alpha of first layer high frequency coefficientFirst layer high frequency coefficient2, threshold correction factor alpha of second layer high frequency coefficientSecond layer high frequency coefficient0.4, threshold correction factor alpha of high-frequency coefficient of third layerThird layer high frequency coefficient0.1, threshold correction factor alpha of third layer low frequency coefficientThird layer low frequency coefficient＝0.03。
And 3.3, comparing the amplitude value of each layer of wavelet coefficient of the carrier audio and the hidden information with a corresponding threshold, if the amplitude value of the wavelet coefficient is larger than the threshold, retaining the amplitude value, otherwise, the wavelet coefficient is 0, and accordingly screening the corresponding wavelet coefficient. Such as: by passing
And screening the high-frequency coefficient of the first layer of the carrier audio.
Step 4, arranging and regulating the amplitude of the wavelet coefficient of the carrier voice in descending order, which comprises the following steps,
step 4.1, wavelet coefficient x of carrier voice1,H[k]、x2,H[k]、x3,H[k]And x3,L[k]Splicing head and tail into array x [ k ]]And then x [ k ]]Descending order to obtain new array x' [ k ]]Each wavelet coefficient is set at x [ k ]]And x' [ k ]]The position corresponding relation (key 1) of the wavelet coefficient recorded in the carrier voice is x [ k ]]And x' [ k ]]Position index J1.
Step 4.2, regulating the amplitude of the carrier voice wavelet coefficient, including ordering the corrected carrier voice wavelet coefficient
I.e. x' [ k ]]The wavelet coefficient amplitude in (1) is normalized to [ - (2)15-1)，(215-1)]Interval and rounding nearby, wherein
For the rounding operation nearby, X [ k ]]Has a length of LX。
And 4.3, counting the number of the corrected carrier voice wavelet coefficients X [ k ] larger than zero as xp and the number of the corrected carrier voice wavelet coefficients X [ k ] smaller than zero as xn.
Step 5, arranging and regulating the wavelet coefficient amplitude of the hidden information in descending order, specifically realizing the steps as follows,
step 5.1, the wavelet coefficient h of the hidden information1,H[k]、h2,H[k]、h3,H[k]And h3,L[k]Splicing head and tail to put into an array h [ k ]]And h [ k ]]Descending the order to obtain a new array h' [ k ]]Setting each wavelet coefficient at h [ k ]]And h' [ k ]]The position corresponding relation (key 2) of the wavelet coefficient recorded in the hidden information is h [ k ]]And h' [ k ]]Position index J2.
Step 5.2, regulating the amplitude of the wavelet coefficient of the hidden information, including order
I.e. h' [ k ]]The wavelet coefficient amplitude in (1) is normalized to [ - (2)13-1)，(213-1)]Interval and rounding nearby, wherein
For the rounding operation nearby, H [ k ]]Has a length of LH。
And 5.3, counting the number of wavelet coefficients Hk of the corrected hidden information which are more than zero as hp and the number of wavelet coefficients Hk which are less than zero as hn.
Step 6, because the audio information hiding method requires that xp + xn is more than or equal to hp + hn, and when the number of the carrier voice wavelet coefficients not being 0 (namely xp + xn) is equal to the number of the hidden information wavelet coefficients not being 0 (namely hp + hn), the capacity of the embedded hidden information is maximum, the invention provides a method capable of adaptively selecting the capacity according to the task requirement, and the concrete implementation steps are as follows,
and 6.1, when the value of xp + xn is less than hp + hn, gradually increasing the length of the carrier audio, increasing the step length to 100, iteratively repeating the processing related to the carrier audio in the steps 1 to 5 for each increase until the condition that the value of xp + xn is more than or equal to hp + hn is met, and entering the step 7.
And 6.2, when the xp + xn > hp + hn, namely the xp + xn is far greater than the hp + hn, gradually reducing the length of the carrier audio, reducing the step length to 100, and repeating the processing related to the carrier audio until the xp + xn is more than or equal to the hp + hn and the xp + xn > hp + hn is not satisfied any more. For example, the embodiment takes the length of the carrier audio to be reduced when xp + xn > hp + hn +100, and iteratively repeats the processing involving the carrier audio in steps 1 to 5 for each reduction until the condition hp + hn ≦ xp + xn ≦ hp + hn +100 is satisfied, and proceeds to step 7.
Step 7, in order to improve the transparency and robustness of the secret-carrying audio, the hidden information embedding of the invention firstly adopts an indirect mode to embed the hidden information, namely the hidden information is converted into an embedding factor and then hidden in the carrier audio; then, the low order bits of the binary number of the carrier audio are replaced by the binary number of the embedding factor by utilizing the idea of least important replacement method so as to obtain the secret carrying information.
The embodiment preferably replaces the last 5 bits of the carrier audio binary number with binary numbers of the embedded factor, with minimal impact on the carrier voice wavelet coefficients, ensuring transparency.
The concrete implementation steps are as follows,
step 7.1,
Binary system of embedding factor of positive wavelet coefficient
Calculating the binary system of the embedding factor of the positive wavelet coefficient, wherein dec2bin (.) represents the decimal system conversion to the binary system calculation;
wherein the content of the first and second substances,
x [ i ] is the hp item before the carrier voice wavelet coefficient X [ k ] after correction;
hi is the hp item before the wavelet coefficient of the concealed information after correction, and i is the serial number of the positive wavelet coefficient of the concealed information.
Binary system of embedding factor of negative wavelet coefficient
j＝LX-hn-1、...LX，j1＝LH-hn-1、...LHAnd calculating the binary system of the embedding factor of the negative wavelet coefficient.
Wherein the content of the first and second substances,
x [ j ] is hn item after carrier voice wavelet coefficient X [ k ] is corrected, j is serial number;
H[j1]for the wavelet coefficient H k of the corrected hidden information]Post hn item, j1Is a serial number;
because of X [ i]Has a maximum value of H [ i ]]4 times the maximum value, therefore XHp[i]And XHn[j]Has a value of (0, 20)]Within the interval, and the binary of each embedded factor occupies 5 significant bits.
Step 7.2, based on the corrected carrier voice wavelet coefficient X [ k ]]Front hp term X [ i]And the modified carrier voice wavelet coefficient X k]Post hn item X [ j ]]Respectively with Xp[i]＝dec2bin(X[i]) Hp and Xn[j]＝dec2bin(X[j]) Hn, j is 1, 2, and X [ i ] is calculated]Binary X ofp[i]And X [ j ]]Binary X ofn[j]. Mixing Xp[i]Last 5 bit of (A) is replaced by XHp[i]Then binary conversion is carried out to decimal calculation to obtain hp item X 'in front of carrier voice wavelet coefficient containing hidden information'p[i]Is mixing Xn[j]Last 5 bit of (A) is replaced by XHn[j]Then binary conversion is carried out to decimal calculation to obtain hn item X 'after the carrier voice wavelet coefficient containing the hidden information'n[j]。
Step 7.3, the x' k obtained in the step 4.1]The front hp wavelet coefficient of (2) is replaced by X'p[i]X' k is]The post-hn wavelet coefficient of (2) is replaced with X'n[j]X' k is mapped to the position stored in J1]The position of the middle element is restored to the position of the original wavelet coefficient, and the wavelet coefficient X embedded with the hidden information is obtainedinsert[k]。
Step 7.4, utilize
Mixing Xinsert[k]Restores to the order of the original wavelet coefficient values and splits the result into x 'according to the stored lengths of the variables Lx1, Lx2, Lx3 and Lx 4'1,H[k](first layer high frequency coefficient), x'2,H[k](second layer high frequency coefficient), x'3,H[k](third layer high frequency coefficient) and x'3,L[k](third layer low frequency coefficient), then 3 layers of inverse wavelet transform are carried out to obtain the secret audio xinsert[n]。
And 8, contrary to the process of embedding the hidden information, extracting the hidden information firstly extracts an embedding factor from the secret carrying audio and then recovers the wavelet coefficient of the hidden information. The concrete implementation steps are as follows,
step 8.1, the received secret-carrying audio yinsert[n]3-layer wavelet transform is carried out to obtain corresponding wavelet coefficient y1,H[k]、y2,H[k]、y3,H[k]And y3,L[k]Splicing them into a plurality of y [ k ] groups]Then use
Normalizing the amplitude to [ - (2)15-1)，(215-1)]The interval is rounded nearby to obtain the wavelet coefficient Y [ k ] of the regular secret-carrying audio]。
Step 8.2, according to the position corresponding relation stored in J1, Y [ k ]]The largest hp elements in the array are taken out and arranged in descending order to form a new array Yp[i]And the smallest hn elements are taken out and arranged in descending order to form a new array Yn[j]And the values of these elements are converted into binary, then the last 5 bits of each binary element are extracted and converted into the embedding factor YH of the decimal hidden information positive wavelet coefficientp[i]Embedding factor YH of negative wavelet coefficient of sum hidden informationn[j]Namely, the embedding factor in the information extraction process is that even if the secret audio is interfered in the channel transmission process, the variable is very small, and the indirect embedding improves the robustness.
Step 8.3, utilize
Obtaining the extracted hidden information positive wavelet coefficient Hp[i]By using
Obtaining the negative wavelet coefficient H of the extracted hidden informationn[j]Then form the extracted hidden information wavelet coefficient in descending order
Step 8.4, according to the corresponding relation of the original position and the descending position of each wavelet coefficient recorded in the J2, converting the H into the Hdesc[k]Converting into extracted hidden information wavelet coefficient Hextra[k]And H is divided by lengths Lh1, Lh2, Lh3 and Lh4extra[k]Is split into h'1,H[k](first layer high frequency coefficient), h'2,H[k](second layer high frequency coefficient), h'3,H[k](third layer high frequency coefficient) and h'3,L[k](third layer low frequency coefficient), then carrying out 3-layer inverse wavelet transform to obtain hidden information hertra[n]。
In specific implementation, the above processes can be automatically operated by adopting a computer software technology.
It should be understood that the above description of the preferred embodiments is given for clarity and not for any purpose of limitation, and that various changes, substitutions and alterations can be made herein without departing from the spirit and scope of the invention as defined by the appended claims.
Claims (3)
1. A method for hiding high-capacity audio information is characterized in that: wavelet coefficients of carrier audio and hidden information are respectively screened by utilizing an adaptive threshold based on average energy, so that the portable audio has better transparency; according to the characteristics of actual concealed voice, the self-adaptive capacity selection is realized by utilizing the characteristics of wavelet coefficients, and the efficiency of embedding concealed information into carrier audio is improved; in the process of embedding the hidden information, firstly, the hidden information is embedded in an indirect mode, the hidden information is converted into an embedding factor and then is hidden in the carrier audio, and then a plurality of low bits of the binary number of the carrier audio are replaced by the binary number of the embedding factor by utilizing the thought of a least important replacement method, so that the secret-carrying information is obtained;
wavelet coefficients of carrier audio and hidden information are respectively screened by utilizing an adaptive threshold based on average energy, the method is realized as follows,
respectively calculating the average energy of wavelet coefficients of each layer of carrier audio and hidden information;
respectively calculating self-adaptive threshold values of wavelet coefficients of each layer of carrier audio and hidden information according to the average energy and corresponding threshold value correction factors;
comparing the amplitude value of wavelet coefficient of each layer of carrier audio frequency and hidden information with corresponding self-adaptive threshold value, if the amplitude value of wavelet coefficient is greater than threshold value retaining, otherwise the wavelet coefficient is 0 so as to implement screening of corresponding wavelet coefficient,
the adaptive capacity selection is implemented as follows,
after screening the wavelet coefficients of the carrier audio and the hidden information, obtaining a carrier voice wavelet coefficient X [ k ] after statistical correction and a wavelet coefficient H [ k ] of the hidden information after correction through descending order arrangement and normalization, wherein the number of the carrier voice wavelet coefficient X [ k ] after statistical correction larger than zero is xp, the number of the carrier voice wavelet coefficient X [ k ] after statistical correction smaller than zero is xn, the number of the wavelet coefficient H [ k ] after statistical correction larger than zero is hp, and the number of the carrier voice wavelet coefficient H [ k ] after statistical correction smaller than zero is hn;
when xp + xn is less than hp + hn, gradually increasing the length of the carrier audio, and re-processing the carrier audio until the condition that xp + xn is more than or equal to hp + hn is met;
when xp + xn > > hp + hn, gradually reducing the length of the carrier audio, and repeating the processing related to the carrier audio until the conditions that xp + xn ≧ hp + hn and xp + xn > > hp + hn are no longer satisfied.
2. The method for hiding high-capacity audio information according to claim 1, wherein: the covert information embedding process includes the steps of,
1) let X [ k ] k]Has a length of LX，H[k]Has a length of LH；
Let X [ i ] be the hp item before the carrier voice wavelet coefficient X [ k ] after correction, Hi be the hp item before the hidden information wavelet coefficient Hk after correction, i be the serial number of the hidden information positive wavelet coefficient,
binary system for calculating embedding factor of positive wavelet coefficient
Wherein dec2bin (·) denotes decimal conversion to binary computation;
let X [ j ]]For modified carrier speech wavelet coefficient X k]The next hn item, j is a serial number; h [ j ]1]For the wavelet coefficient H k of the corrected hidden information]Post hn item, j1Is a serial number; j is LX-hn-1、...LX，j1＝LH-hn-1、...LH
Binary system for calculating embedding factor of negative wavelet coefficient
2) Calculate X [ i ]]Binary X ofp[i]And X [ j ]]Binary X ofn[j]；
Mixing Xp[i]Last 5 bit of (A) is replaced by XHp[i]Then binary conversion is carried out to decimal calculation to obtain hp item X 'in front of carrier voice wavelet coefficient containing hidden information'p[i]，
Mixing Xn[j]Last 5 bit of (A) is replaced by XHn[j]Then binary conversion is carried out to decimal calculation to obtain hn item X 'after the carrier voice wavelet coefficient containing the hidden information'n[j]；
3) When descending order arrangement is set, the wavelet coefficients of the carrier voice are spliced end to end and put into an array x [ k ]]In, for x [ k ]]The descending order is arranged to obtain a new array x' [ k ]]X' k is]The front hp wavelet coefficient of (2) is replaced by X'p[i]X' k is]The post-hn wavelet coefficient of (2) is replaced with X'n[j]Corresponding x' k according to the stored position]The position of the middle element is restored to the position of the original wavelet coefficient, and the wavelet coefficient X embedded with the hidden information is obtainedinsert[k]；
4) By using
Mixing Xinsert[k]The value of (a) is restored to the order of magnitude of the original wavelet coefficient value, the coefficient is split according to the length to the result, then the carrying-secret audio x is obtained by inverse wavelet transformationinsert[n]。
3. The method for hiding high-capacity audio information according to claim 2, wherein: the extraction of the hidden information corresponds to the hidden information embedding process, comprising the following steps,
1) the received secret-carrying audio yinsert[n]Wavelet transform of 3 layers is carried out to obtain wavelet coefficient y1,H[k]、y2,H[k]、y3,H[k]And y3,L[k]Splicing them end to end into an array y [ k ]]Then use
Normalizing the amplitude to [ - (2)15-1)，(215-1)]The interval is rounded nearby to obtain the wavelet coefficient Y [ k ] of the regular secret-carrying audio]；
2) Corresponding Y [ k ] according to the stored position]The largest hp elements in the array are taken out and arranged in descending order to form a new array Yp[i]And the smallest hn elements are taken out and arranged in descending order to form a new array Yn[j]And the values of these elements are converted into binary, then the last 5 bits of each binary element are extracted and converted into the embedding factor YH of the decimal hidden information positive wavelet coefficientp[i]Embedding factor YH of negative wavelet coefficient of sum hidden informationn[j]；
3) By using
Obtaining the extracted hidden information positive wavelet coefficient Hp[i]By using
Obtaining the negative wavelet coefficient H of the extracted hidden informationn[j]Then form the extracted hidden information wavelet coefficient in descending order
4) According to the corresponding relation of the recorded wavelet coefficients at the original position and the positions arranged in descending order, the H is divided intodesc[k]Converting into extracted hidden information wavelet coefficient Hextra[k]And according to the corresponding length, Hextra[k]Splitting into a first layer high-frequency coefficient h'1,H[k]And a second layer high-frequency coefficient h'2,H[k]And the third layer has a high-frequency coefficient h'3,H[ k ] and third layer Low frequency coefficient h'3,L[k]Then, 3-layer inverse wavelet transform is carried out to obtain hidden information hertra[n]。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910243836.7A CN110010142B (en) | 2019-03-28 | 2019-03-28 | Large-capacity audio information hiding method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910243836.7A CN110010142B (en) | 2019-03-28 | 2019-03-28 | Large-capacity audio information hiding method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110010142A CN110010142A (en) | 2019-07-12 |
| CN110010142B true CN110010142B (en) | 2021-05-04 |
Family
ID=67168651
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910243836.7A Expired - Fee Related CN110010142B (en) | 2019-03-28 | 2019-03-28 | Large-capacity audio information hiding method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110010142B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112364386B (en) * | 2020-10-21 | 2022-04-26 | 天津大学 | Audio tampering detection and recovery method combining compressed sensing and DWT |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2002023776A (en) * | 2000-07-13 | 2002-01-25 | Univ Kinki | Method for identifying speaker voice and non-voice noise in blind separation, and method for specifying speaker voice channel |
| CN101625869A (en) * | 2009-08-11 | 2010-01-13 | 中国人民解放军第四军医大学 | Non-air conduction speech enhancement method based on wavelet-packet energy |
| CN102034478A (en) * | 2010-11-17 | 2011-04-27 | 南京邮电大学 | Voice secret communication system design method based on compressive sensing and information hiding |
| CN102142258A (en) * | 2011-03-31 | 2011-08-03 | 上海第二工业大学 | Wavelet transform and Arnold based adaptive gray-scale watermark embedded method |
| CN104851428A (en) * | 2015-04-15 | 2015-08-19 | 东莞中山大学研究院 | Method for realizing watermark encryption and decryption based on audio signal |
| CN108198563A (en) * | 2017-12-14 | 2018-06-22 | 安徽新华传媒股份有限公司 | A kind of Multifunctional audio guard method of digital copyright protection and content authentication |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10319390B2 (en) * | 2016-02-19 | 2019-06-11 | New York University | Method and system for multi-talker babble noise reduction |
-
2019
- 2019-03-28 CN CN201910243836.7A patent/CN110010142B/en not_active Expired - Fee Related
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2002023776A (en) * | 2000-07-13 | 2002-01-25 | Univ Kinki | Method for identifying speaker voice and non-voice noise in blind separation, and method for specifying speaker voice channel |
| CN101625869A (en) * | 2009-08-11 | 2010-01-13 | 中国人民解放军第四军医大学 | Non-air conduction speech enhancement method based on wavelet-packet energy |
| CN102034478A (en) * | 2010-11-17 | 2011-04-27 | 南京邮电大学 | Voice secret communication system design method based on compressive sensing and information hiding |
| CN102142258A (en) * | 2011-03-31 | 2011-08-03 | 上海第二工业大学 | Wavelet transform and Arnold based adaptive gray-scale watermark embedded method |
| CN104851428A (en) * | 2015-04-15 | 2015-08-19 | 东莞中山大学研究院 | Method for realizing watermark encryption and decryption based on audio signal |
| CN108198563A (en) * | 2017-12-14 | 2018-06-22 | 安徽新华传媒股份有限公司 | A kind of Multifunctional audio guard method of digital copyright protection and content authentication |
Non-Patent Citations (5)
| Title |
|---|
| "Adaptive Digital Audio Steganograpghy Based on integer Wavelet Transform";Ahmad Delforouzi;《Circuits Syst Signal Process》;20080304;全文 * |
| "一种基于混沌和小波变换的大容量音频信息隐藏算法";谭良;《电子学报》;20100831;第38卷(第8期);全文 * |
| "一种大容量的自适应信息隐藏方法";谢建全;《计算机应用》;20070531;第27卷(第5期);全文 * |
| "基于小波变换的语音信息隐藏新方法";吴秋玲;《电子与信息学报》;20160430;第38卷(第4期);全文 * |
| "自适应量化索引调制音频信息隐藏算法";田茂;《电子测量技术》;20180220;第41卷(第4期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110010142A (en) | 2019-07-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| El Safy et al. | An adaptive steganographic technique based on integer wavelet transform | |
| Peng et al. | Image watermarking method in multiwavelet domain based on support vector machines | |
| Kashyap et al. | Image watermarking using 3-level discrete wavelet transform (DWT) | |
| Ramanjaneyulu et al. | Wavelet-based oblivious image watermarking scheme using genetic algorithm | |
| Divecha et al. | Implementation and performance analysis of DCT-DWT-SVD based watermarking algorithms for color images | |
| CN112597509B (en) | Information hiding method and system integrating wavelet and self-encoder | |
| Ernawan et al. | Image watermarking based on integer wavelet transform-singular value decomposition with variance pixels | |
| Xu et al. | Exploring robust and blind watermarking approach of colour images in DWT-DCT-SVD domain for copyright protection | |
| CN110010142B (en) | Large-capacity audio information hiding method | |
| Ramanjaneyulu et al. | An oblivious and robust multiple image watermarking scheme using genetic algorithm | |
| Mingguang et al. | A wav-audio steganography algorithm based on amplitude modifying | |
| CN112712811A (en) | Adaptive quantization mean value desynchronization attack robust audio watermarking method | |
| CN101872466A (en) | Watermark embedding method, and watermark detection method and device | |
| Leelavathy et al. | Oblivious Image Watermarking in Discrete Multiwavelet Domain using QIMM. | |
| Hamsathvani | Image hiding in video sequence based on MSE | |
| Sharma et al. | Robust image watermarking technique using contourlet transform and optimized edge detection algorithm | |
| Naik et al. | Joint Encryption and Compression scheme for a multimodal telebiometric system | |
| Guo et al. | Prediction-based watermarking schemes using ahead/post AC prediction | |
| Fan et al. | ROI-based watermarking scheme for JPEG 2000 | |
| Sarma et al. | An Entropy based Video Watermarking Scheme | |
| CN113766084B (en) | Reversible information hiding method and system for enhancing image smoothness | |
| Taheri | Hybrid and Invisible Digital Image Watermarking Technique Using IWT-DCT and Hopfield Neural Network | |
| Phadikar et al. | Multibit QIM watermarking using M-ary modulation and lifting | |
| Ma et al. | A robust watermarking scheme based on dual quantization of wavelet significant difference | |
| Sukumar et al. | Multi Image-Watermarking scheme based on Framelet and SVD |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20210504 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |