Digital audio embedding and detecting method

Abstract

The invention provides a digital audio embedding and detecting method. The digital audio embedding and detecting method comprises the following steps: generating watermark information according to time and a watermark load equipment code; then framing processing is carried out on the audio to be embedded with the watermark, and secondary wavelet decomposition is carried out; embedding a synchronous code in the front half section of the low-frequency wavelet coefficient, performing discrete cosine transform and singular value decomposition processing on the rear half section of the low-frequency wavelet coefficient, and embedding watermark information; and finally, in a watermark detection stage, performing secondary wavelet decomposition on the attacked audio in the current sliding window through a sliding window method, determining whether the current sliding window contains a synchronous code or not, determining the position of the synchronous code, realizing frame synchronization, and processing the rear half section of the audio frame subjected to frame synchronization after attacking so as to extract watermark information. According to the method, the synchronous code and the watermark information are embedded into the audio at the same time, so that the problems that the watermark embedding position is difficult to position and synchronous attacks such as cutting are difficult to resist in the detection stage are solved, and the watermark detection efficiency and accuracy are improved.

Description

Technical Field

[0001] The present application relates to the technical field of digital audio watermarking, and in particular to a digital audio embedding and detection method. Background Art

[0002] A digital audio watermark is a piece of identification generated or embedded based on digital audio content. This identification can be used to identify the copyright of digital audio content and can also be used to protect the integrity of digital audio content. When an audio watermark is used to identify the copyright of digital audio content, it is added to the digital audio content in an imperceptible and non-separable way during the entire life cycle of digital audio content generation, management, distribution and use. Once a copyright dispute occurs, the copyright of the audio digital content can be proved by simply extracting the embedded watermark information from the digital audio content.

[0003] Synchronous attack is the most challenging attack method in the field of digital audio watermarking. In audio watermarking, the watermark is generally embedded at certain fixed points of the audio signal, and the watermark signal is extracted from these fixed points during watermark detection. The purpose of synchronous attack is to destroy the synchronization point of watermark embedding, so that the embedding point of the watermark is offset and the correct watermark cannot be extracted at the detection end. Due to the continuity of audio on the time axis, operations such as cutting and inserting samples at any position will destroy the synchronization structure. Summary of the invention

[0004] The present invention provides a digital audio embedding and detection method, which embeds synchronization code and watermark information into audio at the same time, solves the problem that the watermark embedding position is difficult to locate in the detection stage and is difficult to resist synchronization attacks such as cropping, and improves the watermark detection efficiency and accuracy.

[0005] The present invention provides a digital audio embedding and detection method, which includes three steps: watermark information generation, watermark embedding, and watermark detection. S1: Generate watermark information according to time and watermark load device encoding; S2: Perform frame processing on the audio to be embedded with watermark, and perform secondary wavelet decomposition DWT; embed synchronization code in the first half of low-frequency wavelet coefficients, perform discrete cosine transform DCT and singular value decomposition SVD processing on the second half of low-frequency wavelet coefficients, and embed watermark information; S3: In the watermark detection stage, firstly perform secondary wavelet decomposition on the audio after attack in the current sliding window through the sliding window method, confirm whether the current sliding window contains synchronization code, and determine the synchronization code position, realize frame synchronization, and then process the second half of the audio frame after frame synchronization after attack to extract watermark information.

[0006] Furthermore, the 64-bit watermark information in step S1 includes 16-bit time information, 8-bit year information and 40-bit device code information. The time information is the number of minutes obtained by subtracting 0:00 on January 1 of the current year from the watermark time, divided by 15 and rounded down; the year information is the difference between the year of the current watermark time and 1970; and the device code information is a hexadecimal device code.

[0007] The present invention provides a digital audio watermark embedding method, which adopts the following technical solution:

[0008] The method comprises the following steps S1: generating watermark information according to time and watermark load device encoding; S2: performing frame processing on the audio to be embedded with watermark, and performing secondary wavelet decomposition DWT; embedding synchronization code in the first half of low-frequency wavelet coefficients, performing discrete cosine transform DCT and singular value decomposition SVD processing on the second half of low-frequency wavelet coefficients, and embedding watermark information; S3: in the watermark detection stage, firstly performing secondary wavelet decomposition on the attacked audio in the current sliding window through the sliding window method, confirming whether the current sliding window contains synchronization code, and determining the synchronization code position, realizing frame synchronization, and then processing the second half of the audio frame after frame synchronization after the attack to extract watermark information.

[0009] Furthermore, the watermark information in step S1 is 64-bit watermark information including 16-bit time information, 8-bit year information and 40-bit device coding information.

[0010] Furthermore, the time information is the number of minutes obtained by subtracting 0:00 on January 1 of the current year from the watermark time, divided by 15 and rounded down; the year information is the difference between the year of the current watermark time and 1970; and the device code information is a hexadecimal device code.

[0011] Furthermore, the operation of embedding the synchronization code in the first half of the low-frequency wavelet coefficients in step S2 specifically includes S204: dividing the first half {c(i)|i∈[1,L]} into L equal length b segment, where L b is the number of bits of the synchronization code, and the length of each segment is L B =floor(L / L b ), embed a synchronization code in each segment; then {c(i)|i∈[1,L B ]} is divided into three segments of equal length, and the energy of each segment is calculated and sorted from large to small to obtain E max 、E mid and E min ; S205: When the currently embedded synchronization code bit b(i) is 1, change c(i) according to the following formula:

[0012]

[0013] in The coefficient α must satisfy

[0014]

[0015] When the currently embedded synchronization code bit b(i) is -1, c(i) is changed as follows:

[0016]

[0017] The requirements for S and coefficient α are the same as above, and finally {c′(i)|i∈[1,L]} is obtained after embedding the synchronization code;

[0018] Furthermore, in step S2, the latter half of the low-frequency wavelet coefficients is subjected to discrete cosine transform DCT and singular value decomposition SVD processing, and the operation of embedding 64-bit watermark information specifically includes the following steps: S206: the latter half {c(i)|i∈[L+1,2L]} is subjected to discrete cosine transform, and the 3000-4500Hz segment to which the human ear is insensitive is taken; S207: the 3000-4500Hz segment is segmented according to the number of watermark information bits 64, and each segment is subjected to singular value decomposition (SVD); S208: the first two singular values ​​S1 and S2 are taken to calculate the coefficients When D has the same parity as the current watermark information bit w(i), S1=S2D / 2; when D has different parity from the current watermark information bit w(i), S1=S2(D+1) / 2; S209: after embedding the watermark information, perform inverse SVD transformation to restore; S210: combine the 3000-4500Hz segment after the watermark information is embedded with the unmodified audio segment after DCT transformation, and perform IDCT transformation to obtain {c′(i)|i∈[L+1,2L]}; S211: perform secondary wavelet decomposition inverse transformation IDWT on {c′(i)|i∈[1,L]} obtained in step S205 and {c′(i)|i∈[L+1,2L]} obtained in step S210 to complete the processing of the current audio frame.

[0019] Furthermore, the synchronization code in step S203 is a quasi-m-sequence with a length of 126, and the synchronization code is formed by connecting two m-sequences with a length of 63 end to end.

[0020] Further, step S3 specifically includes the following steps: S301: Sliding window interception, when the audio A after the attack T ={a T (i)|i∈[1,L a]} When the input length is greater than 4L, the input audio is intercepted with a sliding window of length 4L; S302: the audio A in the current sliding window is intercepted T ={a T (i) |i∈[1,4L]} performs secondary wavelet decomposition (DWT) and takes the low-frequency wavelet coefficients {c T (i)|i∈[1,L]}; S303: replace {c T (i)|i∈[1,L]} is divided into equal length L b segment, where L b is the number of bits of the synchronization code, and the length of each segment is L B =floor(L / L b ); then {c T (i)|i∈[1,L B ]} is divided into three segments of equal length, and the energy of each segment is calculated and sorted from large to small to obtain E T max 、E T mid and E T min ; S304: Determine the synchronization code b according to the following formula T (i) S305: The solved b T Convolve with the embedded synchronization code b. If the convolution value is less than the set threshold, it is considered that there is no synchronization code in the current window, and the sliding window continues to slide forward, repeating steps S302-S305; if the convolution value is greater than the set threshold, it is considered that the starting position of the current sliding window is the watermark embedding position, and the subsequent input audio is framed according to the length of 8L based on this starting point; S306: After confirming the synchronization code, the framed audio A T ={a T (i)|i∈[1,8L]}, the first half, you can choose to repeat steps S302-S305 to confirm whether there is a synchronization code; S307: after confirming the synchronization code, the audio A T ={a T (i) The second half of |i∈[1,8L]} is subjected to secondary wavelet decomposition (DWT) and the low-frequency wavelet coefficients {c T (i)|i∈[1,L]}, then perform DCT transformation, and take the 3000-4500Hz segment, and divide it into segments again according to the number of watermark information bits 64. Perform SVD decomposition on each segment, take the first two singular values ​​S1 and S2, and calculate the coefficients When D is an odd number, the current watermark information bit is 1, and when D is an even number, the current watermark information bit is 0. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1It is a flow chart of a digital audio watermark embedding and detection method of the present invention;

[0022] Figure 2 It is a schematic diagram of 64-bit watermark information generated by a digital audio watermark embedding and detection method of the present invention;

[0023] Figure 3 It is a flow chart of watermark embedding in a digital audio watermark embedding and detection method of the present invention;

[0024] Figure 4 It is a flow chart of watermark detection in a digital audio watermark embedding and detection method of the present invention. DETAILED DESCRIPTION

[0025] The present invention is further described in detail below in conjunction with the accompanying drawings.

[0026] In the process of digital audio watermark detection, the synchronization point is destroyed by synchronization attack, resulting in the inability to correctly detect the watermark information. The present invention embeds the synchronization code and the watermark information into the audio at the same time. The synchronization code can improve the synchronization speed of the detection step and resist synchronization attack. Moreover, since the embedding method of the synchronization code is different from the embedding method of the watermark information, the robustness is further improved.

[0027] Reference Figure 1 , the embodiment of the present invention discloses a flow chart of a digital audio embedding and detection method, and the specific operation steps are as follows:

[0028] S1: Generate watermark information according to the time and watermark payload device encoding;

[0029] S2: The audio to be embedded with watermark is framed and decomposed by two levels of wavelet (DWT); the synchronization code is embedded in the first half of the low-frequency wavelet coefficients, and the discrete cosine transform (DCT) and singular value decomposition (SVD) are performed on the second half of the low-frequency wavelet coefficients to embed the watermark information;

[0030] S3: In the watermark detection stage, the audio in the current sliding window after the attack is first decomposed by secondary wavelet to confirm whether the current sliding window contains a synchronization code and determine the position of the synchronization code to achieve frame synchronization. Then, the second half of the audio frame after the attack after frame synchronization is processed to extract the watermark information.

[0031] In this embodiment, the watermark information in step S1 is 64-bit watermark information including 16-bit time information, 8-bit year information and 40-bit device code information. Figure 2As shown in the figure, it contains 16-bit time information, 8-bit year information and 40-bit device code information. The time information is the number of minutes obtained by subtracting 0:00 on January 1 of the current year from the watermark time, divided by 15 and rounded down; the year information is the difference between the current watermark time and 1970; the device code information is the hexadecimal device code.

[0032] Reference Figure 3 The above step S2 is a digital audio watermark embedding method, which specifically includes two steps.

[0033] The first step is to embed the synchronization code in the first half of the low-frequency wavelet coefficients, which specifically includes the following steps:

[0034] S201: Frame processing, the audio to be embedded with watermark A={a(i)|i∈[1,L a ]} is divided into N frames, where L a is the length of the audio, the frame length is 8L, N = floor (L a / 8L);

[0035] S202: Perform secondary wavelet decomposition (DWT) on each audio frame and obtain low-frequency wavelet coefficients {c(i)|i∈[1,2L]};

[0036] S203: Take the first half of the low-frequency wavelet coefficients {c(i)|i∈[1,L]} and embed them into the synchronization code, and take the second half of the low-frequency wavelet coefficients {c(i)|i∈[L+1,2L]} and embed them into the watermark information;

[0037] It should be noted that the synchronization code is added here to achieve fast synchronization positioning of the audio frame. Otherwise, since the embedding method of the watermark information is relatively complex, in the detection stage, since the embedded information content is unknown, synchronization can only be performed by judging the repetition period, which is slow. Therefore, in this embodiment, the low-frequency wavelet coefficients are divided into two halves, one half is used for fast synchronization, and the other half is used for embedding information, so that the speed and robustness of audio watermark detection can be guaranteed.

[0038] S204: Divide the first half {c(i)|i∈[1,L]} into equal lengths L b segment, where L b is the number of bits of the synchronization code, and the length of each segment is L B =floor(L / L b ), embed a synchronization code in each segment; then {c(i)|i∈[1,L B ]} is divided into three segments of equal length, and the energy of each segment is calculated and sorted from large to small to obtain E max 、E mid and E min ;

[0039] S205: When the currently embedded synchronization code bit b(i) is 1, c(i) is changed according to the following formula:

[0040]

[0041] in The coefficient α must satisfy

[0042]

[0043] When the currently embedded synchronization code bit b(i) is -1, c(i) is changed as follows:

[0044]

[0045] The requirements for S and coefficient α are the same as above, and finally {c′(i)|i∈[1,L]} is obtained after embedding the synchronization code;

[0046] In the second step, the discrete cosine transform DCT and singular value decomposition SVD are performed on the second half of the low-frequency wavelet coefficients. The operation of embedding 64-bit watermark information specifically includes the following steps:

[0047] S206: Perform discrete cosine transform (DCT) on the second half {c(i)|i∈[L+1,2L]}, and take the 3000-4500 Hz segment to which human ears are insensitive;

[0048] The frequency range of 3000-4500Hz is selected for processing here to ensure the auditory effect of the original audio. Therefore, the frequency band to which the human ear is insensitive is selected for operation.

[0049] S207: Divide the 3000-4500 Hz segment into segments according to the number of watermark information bits 64, and perform singular value decomposition (SVD) on each segment;

[0050] S208: Take the first two singular values ​​S1 and S2 and calculate the coefficients When D has the same parity as the current watermark information bit w(i),

[0051] S1=S2D / 2;

[0052] When D has different parity from the current watermark information bit w(i),

[0053] S1=S2(D+1) / 2;

[0054] S209: After embedding the watermark information, perform SVD inverse transformation to restore it;

[0055] S210: Combine the 3000-4500 Hz segment after the watermark information is embedded with the unmodified audio segment after the DCT transformation, and perform IDCT transformation to obtain {c′(i)|i∈[L+1,2L]};

[0056] S211: Perform IDWT transformation on {c′(i)|i∈[1,L]} obtained in step S205 and {c′(i)|i∈[L+1,2L]} obtained in step S210, thereby completing the processing of the current audio frame.

[0057] In this embodiment, the frame length 8L in step S201 is 220000.

[0058] In this embodiment, the synchronization code in step S203 is a quasi-m-sequence with a length of 126, and the synchronization code is formed by connecting two m-sequences with a length of 63 end to end.

[0059] In this embodiment, in step S204, {c(i)|i∈[1,L B ]} is divided into three equal sections, each section is 70 meters long.

[0060] Reference Figure 4 The step S3 of the present invention is a digital audio watermark detection method, which specifically includes the following steps:

[0061] S301: Sliding window capture, when the audio A is attacked T ={a T (i)|i∈[1,L a ]} When the input length is greater than 4L, the input audio is intercepted with a sliding window of length 4L;

[0062] S302: Audio A in the current sliding window T ={a T (i) |i∈[1,4L]} performs secondary wavelet decomposition (DWT) and takes the low-frequency wavelet coefficients {c T (i)|i∈[1,L]};

[0063] S303: {c T (i)|i∈[1,L]} is divided into equal length L b segment, where L b is the number of bits of the synchronization code, and the length of each segment is L B =floor(L / L b ); then {c T (i)|i∈[1,L B ]} is divided into three segments of equal length, and the energy of each segment is calculated and sorted from large to small to obtain E T max、E T mid and E T min ;

[0064] S304: Determine the synchronization code b according to the following formula T (i)

[0065]

[0066] S305: The solved b T Convolve with the embedded synchronization code b. If the convolution value is less than the set threshold, it is considered that there is no synchronization code in the current window, and the sliding window continues to slide forward, repeating steps S302-S305; if the convolution value is greater than the set threshold, it is considered that the starting position of the current sliding window is the watermark embedding position, and the subsequent input audio is framed according to the length of 8L with this as the starting point;

[0067] It should be noted that the threshold here is generally not less than 80% of the number of synchronization code bits in applications. Without loss of generality, in this embodiment, the threshold is set to ceil(126*0.8)=101.

[0068] S306: confirm the audio A framed after the synchronization code T ={a T (i) In the first half of |i∈[1,8L]}, steps S302-S305 may be repeated to confirm whether a synchronization code exists;

[0069] S307: confirm the audio A framed after the synchronization code T ={a T (i) The second half of |i∈[1,8L]} is subjected to secondary wavelet decomposition (DWT) and the low-frequency wavelet coefficients {c T (i)|i∈[1,L]}, then perform DCT transformation, and take the 3000-4500Hz segment, and divide it into segments again according to the number of watermark information bits 64. Perform SVD decomposition on each segment, take the first two singular values ​​S1 and S2, and calculate the coefficients When D is an odd number, the current watermark information bit is 1, and when D is an even number, the current watermark information bit is 0.

[0070] In this embodiment, in step S303, {c T (i)|i∈[1,L B ]} is divided into three equal sections, each section is 70 meters long.

[0071] In this embodiment, the threshold in step S305 is set to 101, and the sliding step of the sliding window is set to 1.

[0072] Although the embodiments of the present invention have been shown and described above, it is to be understood that the above embodiments are exemplary and are not to be construed as limitations of the present invention. A person skilled in the art may change, modify, replace and vary the above embodiments within the scope of the present invention.

Claims

1. A digital audio watermark embedding and detection method, characterized in that: The following steps are involved: S1: Generate watermark information according to the time and watermark payload device encoding; S2: The audio to be embedded with watermark is framed and decomposed by secondary wavelet decomposition DWT; the synchronization code is embedded in the first half of the low-frequency wavelet coefficients, and the discrete cosine transform DCT and singular value decomposition SVD are performed on the second half of the low-frequency wavelet coefficients to embed the watermark information; S3: In the watermark detection stage, the audio in the current sliding window after the attack is first decomposed by secondary wavelet to confirm whether the current sliding window contains a synchronization code and determine the position of the synchronization code to achieve frame synchronization. Then, the second half of the audio frame after the attack after frame synchronization is processed to extract the watermark information.

2. The method for embedding and detecting digital audio watermark according to claim 1, characterized in that: The watermark information in step S1 is 64-bit watermark information including 16-bit time information, 8-bit year information and 40-bit device coding information.

3. The method for embedding and detecting digital audio watermark according to claim 2, characterized in that: The time information is the number of minutes obtained by subtracting 0:00 on January 1 of the current year from the watermark time, divided by 15 and rounded down; the year information is the difference between the year of the current watermark time and 1970; and the device code information is a hexadecimal device code.

4. The method for embedding and detecting digital audio watermark according to claim 1, characterized in that: The operation of embedding the synchronization code in the first half of the low-frequency wavelet coefficients in step S2 specifically includes the following steps: S201: Frame processing, the audio to be embedded with watermark A = {a(i)|i∈[1, L a ]} is divided into N frames, where L a is the length of the audio, the frame length is 8L, N = floor (L a / 8L); S202: Perform secondary wavelet decomposition on each audio frame to obtain low-frequency wavelet coefficients {c(i)|i∈[1,2L]}; S203: Take the first half of the low-frequency wavelet coefficients {c(i)|i∈[1, L]} and embed them into the synchronization code, and take the second half of the low-frequency wavelet coefficients {c(i)|i∈[L+1,2L]} and embed them into the watermark information; S204: Divide the first half {c(i)|i∈[1,L]} into equal length L b segment, where L b is the number of bits of the synchronization code, and the length of each segment is L B =floor(L / L b ), each segment is embedded with a synchronization code; then {c(i)|i∈[1, L B ]} is divided into three segments of equal length, and the energy of each segment is calculated and sorted from large to small to obtain E max 、E mid and E min ; S205: When the currently embedded synchronization code bit b(i) is 1, c(i) is changed according to the following formula: in The coefficient α must satisfy When the currently embedded synchronization code bit b(i) is -1, c(i) is changed as follows: Finally, we get {c′(i)|i∈[1,L]} after embedding the synchronization code.

5. The method for embedding and detecting digital audio watermark according to claim 1, characterized in that: In step S2, the discrete cosine transform DCT and singular value decomposition SVD are performed on the second half of the low-frequency wavelet coefficients, and the operation of embedding 64-bit watermark information specifically includes the following steps: S206: Perform discrete cosine transform on the second half {c(i)|i∈[L+1,2L]}, and take the 3000-4500 Hz segment to which human ears are insensitive; S207: Divide the 3000-4500 Hz segment into segments according to the number of watermark information bits 64, and perform singular value decomposition on each segment; S208: Take the first two singular values ​​S1 and S2 and calculate the coefficients When D has the same parity as the current watermark information bit w(i), S1=S2D / 2; When D has different parity from the current watermark information bit w(i), S1=S2(D+1) / 2; S209: After embedding the watermark information, perform SVD inverse transformation to restore it; S210: combine the 3000-4500 Hz segment after the watermark information is embedded with the unmodified audio segment after DCT transformation, and perform inverse DCT transformation to obtain {c′(i)|i∈[L+1,2L]}; S211: Perform secondary inverse wavelet decomposition transform IDWT on {c′(i)|i∈[1,L]} obtained in step S205 and {c′(i)|i∈[L+1,2L]} obtained in step S210, thereby completing the processing of the current audio frame.

6. The method for embedding and detecting digital audio watermark according to claim 4, characterized in that: The synchronization code in step S203 is a quasi-m-sequence with a length of 126, and the synchronization code is composed of two m-sequences with a length of 63 connected end to end.

7. The method for embedding and detecting digital audio watermark according to claim 1, characterized in that: The step S3 specifically comprises the following steps: S301: Sliding window capture, when the audio A is attacked T ={a T (i)|i∈[1,L a ]} When the input length is greater than 4L, the input audio is intercepted with a sliding window of length 4L; S302: Audio A in the current sliding window T ={a T (i) |i∈[1,4L]} performs secondary wavelet decomposition and takes the low-frequency wavelet coefficients {c T (i)|i∈[1,L]}; S303: {c T (i)|i∈[1,L]} is divided into equal length L b segment, where L b is the number of bits of the synchronization code, and the length of each segment is L B =floor(L / L b ); then {c T (i)|i∈[1,L x ]} is divided into three segments of equal length, and the energy of each segment is calculated and sorted from large to small to obtain E T max 、E T mid and E T min ; S304: Determine the synchronization code b according to the following formula T (i) S305: The solved b T Convolve with the embedded synchronization code b. If the convolution value is less than the set threshold, it is considered that there is no synchronization code in the current window, and the sliding window continues to slide forward, repeating steps S302-S305; if the convolution value is greater than the set threshold, it is considered that the starting position of the current sliding window is the watermark embedding position, and the subsequent input audio is framed according to the length of 8L with this as the starting point; S306: confirm the audio A framed after the synchronization code T ={a T (i) In the first half of |i∈[1,8L]}, steps S302-S305 may be repeated to confirm whether a synchronization code exists; S307: confirm the audio A framed after the synchronization code T ={a T (i) The second half of |i∈[1,8L]} is subjected to secondary wavelet decomposition DWT, and the low-frequency wavelet coefficients {c T (i)|i∈[1,L]}, then perform DCT transformation, and take the 3000-4500Hz segment, and divide it into segments again according to the number of watermark information bits 64. Perform SVD decomposition on each segment, take the first two singular values ​​S1 and S2, and calculate the coefficients When D is an odd number, the current watermark information bit is 1, and when D is an even number, the current watermark information bit is 0.

Images