KR101333825B1 - Apparatus and method of audio watermarking, and, apparatus and method of audio watermark detecting - Google Patents

Abstract

The present invention relates to an audio watermark insertion device, method thereof, audio watermark detection device and a method therefor. The invention determines a frequency range having maximum energy and inserts the watermark in a QIM method using an adaptive step size based on response amplitude of the frequency range. [Reference numerals] (100) First band selection unit;(200) Watermark insertion unit;(300) Signal generation unit;(AA) Start;(BB) End

Description

Apparatus and method of audio watermarking, and, apparatus and method of audio watermark detecting}

The present invention relates to an audio watermark embedding apparatus and method, and an audio watermark detecting apparatus and method. More particularly, the present invention relates to an audio watermark embedding apparatus and method. An audio watermark embedding apparatus and method for embedding a watermark in a QIM method using a step size, and an audio watermark detecting apparatus and method.

With the rapid development of the Internet and multimedia technologies, the social demand for the production and distribution of digital contents has increased.

Requests for copyright protection are increasing day by day, and techniques for protecting the copyright of multimedia works are being researched to prevent illegal copying and distribution of digital contents.

Among them, watermark technology can protect the ownership of multimedia data by inserting copyright information into digital image, audio, and video signals that are difficult to detect through human eyes or ears for the purpose of copyright protection of digital contents. It is also known as a technology that can prevent illegal copying of indiscriminate data. In particular, audio watermarking must be robust to signal processing such as compression, sampling, and at the same time maintain audio quality so that it is inaudible to human hearing.

However, conventional watermarking techniques have not shown sufficient robustness in signal processing. This is because it is difficult to select the best transform domain in which the watermark is to be inserted. Therefore, there is a need for a method that is robust to signal processing and that is easy to insert and detect a watermark.

In order to solve the above problem, Korean Patent Laid-Open Publication No. 10-2011-0014871 discloses a method of converting an audio signal into a frequency domain and inserting a watermark at a peak of a frequency band including maximum energy. The present invention has the effect of being robust to signal processing while maintaining the sound quality of the audio signal even after the watermark is inserted into the digital audio signal by inserting the watermark at the peak of the maximum frequency band.

However, Korean Patent Laid-Open Publication No. 10-2011-0014871 is a technology for private watermarking in which an original audio signal is required when watermark is detected, and there is a problem in that a watermark cannot be detected in the absence of the original audio signal.

Republic of Korea Patent Publication No. 10-2011-0014871

SUMMARY OF THE INVENTION The present invention has been made to solve the above problems, and an object of the present invention is to provide an audio watermark embedding apparatus and method thereof capable of detecting a watermark without an original audio signal, and a watermark detecting apparatus and method thereof.

Another object of the present invention is to provide an audio watermark embedding apparatus and method, and a watermark detecting apparatus and method that are more resistant to signal attack than in the related art.

An audio watermark embedding apparatus according to an aspect of the present invention comprises: a first band selecting unit configured to select a frequency band having a maximum energy among a plurality of frequency bands separated by applying a discrete wavelet transform (DWT) to an original audio signal; A frequency band signal generated by applying an inverse discrete wavelet transform (IDWT) and a fast fourier transform (FFT) to the selected frequency band signal is divided into a plurality of frames, and a QIM (Qunatization Index Modulation) is applied to each frame sample. A watermark insertion unit for embedding the watermark in a manner; And applying an Inverse Fast Fourier Transform (IFFT) to the frame signal having the watermark embedded therein, and applying an Inverse Discrete Wavelet Transform (IDWT) to a signal of a frequency band not selected by the first band selector. And a signal generator for generating the audio signal.

In this case, the first band selector may include: a first DWT converter configured to apply DWT to the original audio signal and divide the signal into a plurality of frequency bands; And a first selector configured to select a frequency band having the maximum energy among the plurality of frequency bands.

The watermark embedding unit may include: a first IDWT converter configured to reconstruct the band signal by applying an IDWT to the signal of the selected band; A first frame converter applying an FFT to the reconstructed signal and separating the FFT-converted signal into a plurality of frames; A first step size determination unit determining a step size for each frame by using a frequency response size of each frame of the separated frame; And a QIM unit for inserting a watermark into the samples for each frame by the QIM method by using the step size corresponding to each frame.

The watermark embedding unit may further include a first threshold value setting unit configured to set a signal threshold value of each frame by using a step size corresponding to each frame, and the QIM unit may have a frequency response size of each frame. The watermark may be inserted only for a sample having a signal threshold greater than or equal to.

The signal generator may include a first IFFT converter configured to reconstruct a signal in a spatial domain by applying an IFFT to the frame signal having the watermark embedded therein; A second IDWT converter configured to reconstruct a signal in a spatial domain by applying IDWT to a signal in a frequency band not selected by the first band selector; And a reconstruction unit for generating a watermark-embedded audio signal using the signals reconstructed by the first IFFT converter and the second IDWT converter.

An audio watermark embedding method according to another aspect of the present invention includes the steps of selecting a frequency band having the maximum energy among a plurality of frequency bands separated by applying a discrete wavelet transform (DWT) to the original audio signal; A frequency band signal generated by applying an inverse discrete wavelet transform (IDWT) and a fast fourier transform (FFT) to the selected frequency band signal is divided into a plurality of frames, and a QIM (Qunatization Index Modulation) is applied to each frame sample. Embedding the watermark in a manner; And applying an Inverse Fast Fourier Transform (IFFT) to the frame signal having the watermark embedded therein, and applying an Inverse Discrete Wavelet Transform (IDWT) to a signal of a frequency band not selected in the band selection step. And a signal generation step of generating a signal.

In this case, the step of selecting a frequency band to insert the watermark, the step of applying a DWT to the original audio signal to be divided into a plurality of frequency bands; And selecting a frequency band having the maximum energy among the plurality of frequency bands.

The embedding of the watermark may include reconstructing the band signal by applying IDWT to the signal of the selected band; Applying an FFT to the reconstructed signal, and separating the FFT transformed signal into a plurality of frames; Determining a step size for each frame using the frequency response magnitude of each frame of the separated frame; And inserting a watermark in the QIM method into the samples for each frame by using the step size corresponding to each frame.

The inserting of the watermark may further include setting a signal threshold of each frame by using a step size corresponding to each frame, and inserting the watermark by the QIM method. For example, the watermark may be inserted only for a sample whose frequency response size is greater than or equal to the signal threshold of each frame.

The generating of the watermark-embedded audio signal may include reconstructing a signal of a spatial domain by applying an IFFT to the watermarked frame signal; Reconstructing a signal of a spatial domain by applying IDWT to a signal of an unselected frequency band in selecting the maximum energy frequency band; And generating a watermark-embedded audio signal using the reconstructed signals through applying the IFFT and applying the IDWT.

The audio watermark detection apparatus according to another aspect of the present invention is a second band for selecting a frequency band having the maximum energy among a plurality of frequency bands separated by applying a DWT (Discrete Wavelet Transform) to the watermark-embedded audio signal Selection unit; And splitting the signal of the frequency band generated by applying an inverse discrete wavelet transform (IDWT) and a fast fourier transform (FFT) to the selected frequency band signal into a plurality of frames, and determining the Euclidean distance for each frame sample. It includes a watermark detection unit for detecting the watermark by using.

In this case, the second band selector may include: a second DWT converter configured to apply a DWT to the audio signal into which the watermark is inserted and to divide the frequency band into a plurality of frequency bands; And a second selector configured to select a frequency band having the maximum energy among the plurality of frequency bands.

The watermark detection unit may further include: a third IDWT converter configured to reconstruct the band signal by applying an IDWT to the signal of the selected band; A second frame converter configured to apply an FFT to the reconstructed signal and separate the FFT-converted signal into a plurality of frames; A second step size determination unit determining a step size for each frame by using a frequency response size of each frame of the separated frame; And a detector configured to detect the watermark using the step size and Euclidean distance corresponding to each frame.

In this case, the watermark detection unit further includes a second threshold value setting unit that sets a signal threshold value of each frame by using a step size corresponding to each frame, and the detection unit has a frequency response size of each frame. The watermark can be detected only for samples that are equal to or greater than the signal threshold of.

According to another aspect of the present invention, an audio watermark detection method includes selecting a frequency band having a maximum energy among a plurality of frequency bands separated by applying a discrete wavelet transform (DWT) to an audio signal having a watermark embedded therein; And splitting the signal of the frequency band generated by applying an inverse discrete wavelet transform (IDWT) and a fast fourier transform (FFT) to the selected frequency band signal into a plurality of frames, and determining the Euclidean distance for each frame sample. Detecting the watermark using the same.

In this case, the step of selecting a frequency band for detecting the watermark, the step of separating the plurality of frequency bands by applying a DWT to the audio signal embedded watermark; And selecting a frequency band having the maximum energy among the plurality of frequency bands.

The detecting of the watermark may include reconstructing the band signal by applying IDWT to the signal of the selected band; Applying an FFT to the reconstructed signal, and separating the FFT transformed signal into a plurality of frames; Determining a step size for each frame using the frequency response magnitude of each frame of the separated frame; And detecting the watermark using the step size and Euclidean distance corresponding to each frame.

In this case, the detecting of the watermark may further include setting a signal threshold value of each frame using the step size corresponding to each frame, wherein the step size and Euclid correspond to each frame. In the detecting of the watermark using the dian distance, the watermark may be detected only for a sample whose frequency response size is greater than or equal to the signal threshold of each frame.

An audio watermark embedding apparatus and method thereof, and a watermark detecting apparatus and method thereof according to a preferred embodiment of the present invention have the effect of detecting a watermark without an original audio signal.

In addition, by inserting a watermark in a QIM method using an adaptive step size for each frame of the audio signal converted into the frequency domain, it is more effective to attack the signal.

1 is a block diagram showing an audio watermark embedding apparatus according to an embodiment of the present invention;
2 is a block diagram illustrating a first band selecting unit according to an embodiment of the present invention;
3 is a diagram illustrating a process of separating an original audio signal into a plurality of frequency bands by a first DWT converter according to an embodiment of the present invention;
4 is a block diagram showing a watermark insertion unit according to an embodiment of the present invention;
5 is an indicator showing the robustness of the watermark when the watermark is inserted according to an embodiment of the present invention;
6 is a block diagram showing a signal generator according to an embodiment of the present invention;
7 is a flowchart illustrating an audio watermark embedding method according to another embodiment of the present invention;
8 illustrates a process of separating an original audio signal into a plurality of frequency bands by DWT conversion according to another embodiment of the present invention;
9 is an index showing the robustness of a watermark when inserting a watermark according to another embodiment of the present invention;
10 is a block diagram showing an audio watermark detection apparatus according to another embodiment of the present invention;
11 is a block diagram illustrating a second band selecting unit according to another embodiment of the present invention;
12 illustrates a process of separating an audio signal having a watermark into a plurality of frequency bands by a second DWT converter according to another embodiment of the present invention;
13 is a block diagram showing a watermark detection unit according to another embodiment of the present invention;
14 is a flowchart illustrating an audio watermark detection method according to another embodiment of the present invention; and
FIG. 15 illustrates a process of separating an audio signal having a watermark into a plurality of frequency bands by DWT conversion according to another embodiment of the present invention.

BRIEF DESCRIPTION OF THE DRAWINGS The present invention is capable of various modifications and various embodiments, and specific embodiments are illustrated in the drawings and described in detail in the detailed description. It should be understood, however, that it is not intended to be limited to the specific embodiments of the invention but includes all modifications, equivalents, and alternatives falling within the spirit and scope of the invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, the present invention will be described in detail with reference to the accompanying drawings.

The terms first, second, etc. may be used to describe various components, but the components should not be limited by the terms. The terms are used only for the purpose of distinguishing one component from another.

The terminology used herein is for the purpose of describing particular example embodiments only and is not intended to be limiting of the present invention. Singular expressions include plural expressions unless the context clearly indicates otherwise. In this application, the terms "comprise" or "have" are intended to indicate that there is a feature, number, step, operation, component, part, or combination thereof described in the specification, and one or more other features. It is to be understood that the present invention does not exclude the possibility of the presence or the addition of numbers, steps, operations, components, components, or a combination thereof.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings.

1 is a block diagram showing an audio watermark embedding apparatus according to an embodiment of the present invention.

As shown in FIG. 1, an audio watermark embedding apparatus according to a preferred embodiment applicable to the present invention includes a frequency having a maximum energy among a plurality of frequency bands separated by applying a discrete wavelet transform (DWT) to an original audio signal. A first band selecting unit 100 for selecting a band; A frequency band signal generated by applying an inverse discrete wavelet transform (IDWT) and a fast fourier transform (FFT) to the selected frequency band signal is divided into a plurality of frames, and a QIM (Qunatization Index Modulation) is applied to each frame sample. A watermark inserting unit 200 for inserting a watermark in a manner; And applying an Inverse Fast Fourier Transform (IFFT) to the frame signal having the watermark embedded therein, and applying an Inverse Discrete Wavelet Transform (IDWT) to a signal of a frequency band not selected by the first band selector 100. And a signal generator 300 for generating an audio signal with a mark inserted therein.

The first band selector 100 selects a frequency band having the maximum energy among a plurality of frequency bands separated by applying DWT to the original audio signal. This will be described in detail with reference to FIG.

2 is a block diagram illustrating a first band selecting unit according to an embodiment of the present invention.

As shown in FIG. 2, in a preferred embodiment applicable to the present invention, the first band selector 100 includes a first DWT converter 110 for applying a DWT to an original audio signal and separating the signal into a plurality of frequency bands; And a first selector 120 that selects a frequency band having the maximum energy among the plurality of frequency bands.

The first DWT converter 110 separates the original audio signal into a plurality of frequency bands as shown in FIG. 3 by applying DWT to the original audio signal. In FIG. 3, the original audio signal is divided into six frequency bands, but the present invention is not limited thereto. The DWT transformation has a characteristic that there is no blocking phenomenon by transforming the entire signal and that both frequency information and position information can be expressed due to the characteristics of the transfer function. The DWT also allows for multiple resolution analysis. DWT having such characteristics is applied to various fields such as signal processing.

The first selector 120 selects a frequency band having the maximum energy among the plurality of frequency bands. In order to insert a watermark that is robust to signal processing, it is important to insert a watermark in the main part of the signal, and to design a serious sound quality degradation when attempting to remove the watermark. In a preferred embodiment applicable to the present invention, the first selector 120 selects a frequency band having the maximum energy and inserts a watermark into the selected frequency band to provide a watermarking apparatus that is robust to a signal processing attack. do. In another embodiment, when the energy magnitudes of two frequency bands are the same or difficult to compare, the frequency band including the pitch among the two frequency bands may be selected as the watermark embedding band.

The watermark inserting unit 200 divides the signal of the frequency band generated by applying IDWT and FFT to the frequency band signal selected by the first band selecting unit 100 into a plurality of frames, and samples each frame. Insert a watermark in the QIM method. This will be described in detail with reference to FIG.

4 is a block diagram illustrating a watermark insertion unit according to an embodiment of the present invention.

As shown in FIG. 4, in a preferred embodiment applicable to the present invention, the watermark inserting unit 200 includes: a first IDWT converting unit 210 reconstructing the band signal by applying IDWT to the signal of the selected band; A first frame converter 220 applying an FFT to the reconstructed signal and separating the FFT-converted signal into a plurality of frames; A first step size determiner 230 for determining a step size for each frame by using a frequency response size of each frame of the separated frame; A first threshold setting unit 235 for setting a signal threshold of each frame by using a step size corresponding to each frame; And a QIM unit 240 for inserting a watermark in a QIM method into the samples for each frame by using the step size corresponding to each frame.

The first IDWT converter 210 reconstructs the band signal by applying IDWT to a signal of a band selected as a frequency band into which a watermark is to be inserted. More specifically, before inserting a watermark in the selected frequency band, IDWT is applied to the signal and reconstructed into a signal in the spatial domain.

The first frame converter 220 converts the signal into a frequency domain by applying an FFT to the reconstructed signal. The present invention relates to an audio watermark embedding apparatus for watermarking by applying a QIM method in a frequency domain. In an embodiment applicable to the present invention, the first frame converter 220 performs an FFT on a signal reconstructed into a spatial domain. Apply watermark to insert watermark in frequency domain.

In addition, the first frame converter 220 separates the FFT-converted signal into a plurality of frames. The present invention relates to an audio watermark embedding apparatus for inserting a watermark (1) bit into a sample of each of the separated frames. In an embodiment applicable to the present invention, the first frame converter 220 includes a rectangular window. By using it can be separated into a frame unit containing M (M is a natural number) samples without overlap.

The first step size determiner 230 determines an adaptive step size for each of the separated frames using the frequency response size of each frame. In a preferred embodiment applicable to the present invention, the first step size determiner 230 may determine the step size for each frame using Equation 1 below.

는 k번째 프레임의 스텝사이즈,

는 k번째 프레임의 주파수 응답 크기,

는 신호 전체에 대해 고정된 파라미터)
(

Is the step size of the kth frame,

Is the magnitude of the frequency response of the kth frame,

Is a fixed parameter for the entire signal)

When inserting a watermark using the QIM method, increasing the quantization step size lowers the sound quality of the watermarked signal, while decreasing the step size decreases the watermark's robustness. In other words, it is necessary to determine the appropriate quantization step size. However, due to the characteristics of the watermark embedding method using the QIM method, the quantization step size appropriate for each frame differs according to the frequency response size of each frame. For example, in the case of a frame having a relatively large frequency response size, even if a watermark is inserted by applying a step size larger than another frame, the sound quality of the signal does not significantly decrease. Therefore, in an embodiment applicable to the present invention, the first step size determiner 230 may determine an adaptive step size according to the maximum frequency response size of each frame. At this time, in an embodiment applicable to the present invention, the first stabilizer determiner 230 may determine the step size for each frame using Equation (1).

)는 각 프레임 별 스텝사이즈의 크기를 조절하는 값으로, 원 오디오 신호의 특성에 따라 상기 고정된 파라미터(

)의 크기는 변경될 수 있다. 본 발명에 대한 일 실시예에서는 상기 고정된 파라미터(

)값으로 10을 적용하였다. 이에 대한 워터마크 강인성 정도는 이하 도 5를 통해 상세히 설명한다.
Fixed parameter of Equation 1

) Is a value that adjusts the size of the step size for each frame, and the fixed parameter (

) Can be changed in size. In one embodiment of the present invention the fixed parameter (

A value of 10 was applied. The degree of watermark robustness for this will be described in detail with reference to FIG. 5.

The first threshold value setting unit 235 may set a signal threshold value of each frame by using a step size corresponding to each frame. The QIM unit 240 of the present invention inserts a watermark using a QIM method using respective quantizers corresponding to m = {0,1}, wherein a signal sample having a specific frequency response size is a watermark embedding method of a QIM method. Can be quantized to negative.

In the frequency domain signal having a magnitude value and a phase value, since the magnitude value should always have a positive value, a signal threshold may be set to prevent the magnitude value from being negative. More specifically, in an embodiment applicable to the present invention, when the first threshold value setting unit 235 sets a signal threshold value of a predetermined size for each frame and then inserts a watermark through the QIM unit 240, A watermark may be inserted only for samples whose frequency response size is greater than or equal to the signal threshold of each frame, thereby preventing the frequency quantization from being negative. Hereinafter, the QIM unit 240 will be described in detail.

으로 설정하였다. 이에 대한 워터마크 강인성 정도는 이하 도 5를 통해 상세히 설명한다.
At this time, in an embodiment of the present invention, the signal threshold of each frame is determined.

Set to. The degree of watermark robustness for this will be described in detail with reference to FIG. 5.

The QIM unit 240 inserts a watermark in a QIM method into the samples for each frame by using the step size corresponding to each frame. The method of inserting the watermark by the QIM unit 240 using the QIM method is as follows.

을 삽입하기 위해, [

,

]에 걸친 의사랜덤 균일 분포를 가지는 디더 벡터 d[k,0]를 만든다. 이어서 상기 QIM부(240)는 하기 수학식 2를 이용하여 디더 벡터 d[k,1]를 만든다.First, the QIM unit 240 has a watermark of one bit in each frame of a sample.

To insert it, [

,

Produces a dither vector d [k, 0] with a pseudorandom uniform distribution over. Subsequently, the QIM unit 240 generates a dither vector d [k, 1] using Equation 2 below.

에 대한 상기 디더 벡터 d(m)을 이용하여 신호를 양자화시키는 방법으로 하기 수학식 3과 같이 워터마크가 삽입된 신호 s(x;m)를 만든다. 이때, 상기 QIM부(240)는 상기 제1 스텝사이즈 결정부(230)로부터 결정된 각 프레임별로 적응적인 스텝사이즈를 이용하여 워터마크를 삽입한다.Watermark

By using the dither vector d (m) for quantizing a signal, a watermark-embedded signal s (x; m) is generated as shown in Equation 3 below. In this case, the QIM unit 240 inserts a watermark using an adaptive step size for each frame determined by the first step size determiner 230.

인 균일 스칼라 양자화기)(q (·) is the step size

Uniform scalar quantizer)

In a preferred embodiment applicable to the present invention, the QIM unit 240 may insert a watermark only for a sample whose frequency response size of the sample for each frame is greater than or equal to the signal threshold of each frame. As described above, since the magnitude of the frequency response of the signal should not be negative through QIM quantization, the watermark is inserted only for samples having a magnitude greater than or equal to the signal threshold of each frame. Negative quantization can be prevented.

)는 10의 값을 갖는다.), 상기 각 프레임의 신호 문턱값이

일 때의 워터마크 강인성을 나타낸 지표이다.5 is an index showing the robustness of a watermark when inserting a watermark according to an embodiment of the present invention. More specifically, the step size for each frame is determined by using Equation 1 above, where a fixed parameter (

) Has a value of 10.), and the signal threshold of each frame

It is an index indicating the watermark toughness when.

As shown in FIG. 5, the inserted watermark according to the preferred embodiment applicable to the present invention shows that the BER is less than about 5% for most signal attacks, and thus, it is confirmed that the watermark is robust to the attack.

The signal generator 300 applies an IFFT to the frame signal having the watermark embedded therein, and applies an IDWT to a signal having a frequency band not selected by the first band selector 100 so that the watermark has been embedded. Create Hereinafter, a detailed description will be given with reference to FIG. 6.

6 is a block diagram illustrating a signal generator according to an embodiment of the present invention.

As shown in FIG. 6, in a preferred embodiment applicable to the present invention, the signal generator 300 may include a first IFFT converter configured to reconstruct a signal of a spatial domain by applying an IFFT to the frame signal having the watermark embedded therein. 310); A second IDWT converter 320 reconstructing a signal in a spatial domain by applying IDWT to a signal in a frequency band not selected by the first band selector 100; And a reconstruction unit 330 for generating an audio signal including a watermark by using the signals reconstructed by the first IFFT converter 310 and the second IDWT converter 320.

The first IFFT converter 310 reconstructs a signal of a spatial domain by applying an IFFT to a signal of a frame having a watermark embedded therein. In addition, the second IDWT converter 320 reconstructs the signal in the spatial domain by applying IDWT to a signal of a frequency band not selected by the first band selector 100.

Thereafter, the reconstructor 330 generates an audio signal with a watermark embedded using the signals reconstructed by the first IFFT converter 310 and the second IDWT converter 320. In more detail, IFFT is applied to a signal in the frequency domain where the watermark is embedded, and reconstructed into a signal in the spatial domain. The signal is DWT-converted by the first DWT converter 110 and the watermark is not inserted. IDWT is applied to reconstruct the signal in the spatial domain. As described above, the reconstruction unit 330 generates an audio signal having a watermark embedded therein using each signal reconstructed as a signal of a spatial domain.

7 is a flowchart illustrating an audio watermark embedding method according to another embodiment of the present invention.

As shown in FIG. 7, an audio watermark embedding method according to a preferred embodiment applicable to the present invention includes: separating DW into a plurality of frequency bands by applying DWT to an original audio signal; Selecting a frequency band having the maximum energy among the plurality of frequency bands; Reconstructing the band signal by applying IDWT to the signal of the selected band; Applying an FFT to the reconstructed signal, and separating the FFT transformed signal into a plurality of frames; Determining a step size for each frame using the frequency response magnitude of each frame of the separated frame; Setting a signal threshold of each frame using the step size corresponding to each frame; Inserting a watermark in the QIM method into the samples for each frame by using the step size corresponding to each frame; Reconstructing a signal of a spatial domain by applying an IFFT to the frame signal having the watermark embedded therein; Reconstructing a signal in a spatial domain by applying IDWT to a signal in a frequency band not selected as a frequency band into which a watermark is to be inserted; And generating a watermark-embedded audio signal using the reconstructed signals through applying the IFFT and applying the IDWT.

The DWT is applied to the original audio signal and separated into a plurality of frequency bands (S110). At this time, the original audio signal is separated into a plurality of frequency bands as shown in FIG. In the case of FIG. 8, the original audio signal is divided into six frequency bands, but the present invention is not limited thereto. The DWT transformation has a characteristic that there is no blocking phenomenon by transforming the entire signal and that both frequency information and position information can be expressed due to the characteristics of the transfer function. The DWT also allows for multiple resolution analysis. DWT having such characteristics is applied to various fields such as signal processing.

The frequency band having the maximum energy among the plurality of frequency bands separated in step S110 is selected (S120). In order to insert a watermark that is robust to signal processing, it is important to insert a watermark in the main part of the signal, and to design a serious sound quality degradation when attempting to remove the watermark. According to a preferred embodiment of the present invention, a watermarking method that is robust to signal processing attack is provided by selecting a frequency band having the maximum energy in step S120 and inserting a watermark in the selected frequency band. In another embodiment, when the energy magnitudes of two frequency bands are the same or difficult to compare, the frequency band including the pitch among the two frequency bands may be selected as the watermark embedding band.

The band signal is reconstructed by applying IDWT to the signal of the band selected in step S120 (S130). More specifically, before inserting a watermark in the selected frequency band, IDWT is applied to the signal and reconstructed into a signal in the spatial domain.

The FFT is applied to the reconstructed signal in step S130 to convert the signal into a frequency domain signal, and the FFT-converted signal is separated into a plurality of frames (S140).

The present invention relates to an audio watermark embedding method for watermarking by applying a QIM method in a frequency domain. In an embodiment applicable to the present invention, an FFT is applied to a signal of a spatial domain reconstructed in step S130 to determine a frequency domain. By converting to a signal, a watermark can be inserted in the frequency domain.

In addition, the FFT transformed signal is separated into a plurality of frames. The present invention relates to an audio watermark embedding method for inserting one (1) bit of watermark into a sample of each of the separated frames. In an embodiment applicable to the present invention, the step S140 is performed without overlap by using a rectangular window. (M is a natural number) and may be separated into frame units including samples.

The adaptive step size is determined for each frame separated in the step S140 by using the frequency response size of each frame (S150). In a preferred embodiment applicable to the present invention, the step S150 may determine the step size for each frame using Equation 4 below.

는 k번째 프레임의 스텝사이즈,

는 k번째 프레임의 주파수 응답 크기,

는 신호 전체에 대해 고정된 파라미터)
(

Is the step size of the kth frame,

Is the magnitude of the frequency response of the kth frame,

Is a fixed parameter for the entire signal)

When inserting a watermark using the QIM method, increasing the quantization step size lowers the sound quality of the watermarked signal, while decreasing the step size decreases the watermark's robustness. In other words, it is necessary to determine the appropriate quantization step size. However, due to the characteristics of the watermark embedding method using the QIM method, the quantization step size appropriate for each frame differs according to the frequency response size of each frame. For example, in the case of a frame having a relatively large frequency response size, even if a watermark is inserted by applying a step size larger than another frame, the sound quality of the signal does not significantly decrease. Therefore, in the embodiment applicable to the present invention, the step S150 may determine an adaptive step size according to the maximum frequency response size of each frame. At this time, in an embodiment applicable to the present invention, the step size for each frame may be determined using Equation 4.

)는 각 프레임 별 스텝사이즈의 크기를 조절하는 값으로, 원 오디오 신호의 특성에 따라 상기 고정된 파라미터(

)의 크기는 변경될 수 있다. 본 발명에 대한 일 실시예에서는 상기 고정된 파라미터(

) 값으로 10을 적용하였다. 이에 대한 워터마크 강인성 정도는 이하 도 9를 통해 상세히 설명한다.
Fixed parameter of Equation 4

) Is a value that adjusts the size of the step size for each frame, and the fixed parameter (

) Can be changed in size. In one embodiment of the present invention the fixed parameter (

10 was applied. The degree of watermark robustness for this will be described in detail with reference to FIG. 9.

In an embodiment applicable to the present invention, the signal threshold of each frame may be set using the step size corresponding to each frame (S160). The watermark embedding method according to another embodiment of the present invention inserts a watermark using a QIM method using respective quantizers corresponding to m = {0,1}, wherein signal samples having a specific frequency response size are smaller than the QIM method. It can be quantized to a negative number by the watermark embedding method of.

In the frequency domain signal having a magnitude value and a phase value, the magnitude value should always have a positive value, so that the magnitude value can be prevented from being negative by setting the signal threshold. More specifically, in an embodiment applicable to the present invention, a signal threshold value of a predetermined size is set for each frame, and when a watermark is inserted thereafter, the watermark is inserted only for a simpler that is greater than or equal to the signal threshold value of each frame. The phenomenon of being quantized to this negative number can be prevented. The detailed description will be described in detail in the watermark embedding step.

으로 설정하였다. 이에 대한 워터마크 강인성 정도는 이하 도 9를 통해 상세히 설명한다.
In an embodiment of the present invention, the signal threshold of each frame is

Was set. The degree of watermark robustness for this will be described in detail with reference to FIG. 9.

A watermark is inserted into the sample for each frame by the QIM method using the step size corresponding to each frame (S170). A method of inserting a watermark using the QIM method in step S170 is as follows.

을 삽입하기 위해, [

,

]에 걸친 의사랜덤 균일 분포를 가지는 디더 벡터 d[k,0]를 만든다. 이어서 하기 수학식 5를 이용하여 디더 벡터 d[k,1]를 만든다.First, a bit of watermark

To insert it, [

,

Produces a dither vector d [k, 0] with a pseudorandom uniform distribution over. Subsequently, the dither vector d [k, 1] is generated by using Equation 5 below.

에 대한 상기 디더 벡터 d(m)을 이용하여 신호를 양자화시키는 방법으로 하기 수학식 6과 같이 워터마크가 삽입된 신호 s(x;m)를 만든다. 이때, 상기 S160 단계에서 결정된 각 프레임별로 적응적인 스텝사이즈를 이용하여 워터마크를 삽입한다.Watermark

By using the dither vector d (m) for quantizing the signal, a watermark-embedded signal s (x; m) is generated as shown in Equation 6 below. At this time, the watermark is inserted using the adaptive step size for each frame determined in step S160.

(q (·) is the step size Uniform scalar quantizer)

In a preferred embodiment applicable to the present invention, the watermark may be inserted only for a sample whose frequency response magnitude of the sample for each frame is greater than or equal to the signal threshold of each frame. As described above, since the magnitude of the frequency response of the signal should not be negative through QIM quantization, the watermark is inserted only for samples having a magnitude greater than or equal to the signal threshold of each frame. Negative quantization can be prevented.

)는 10의 값을 갖는다,), 상기 각 프레임의 신호 문턱값이

일 때의 워터마크 강인성을 나타낸 지표이다.9 is an index showing the robustness of a watermark when inserting a watermark according to an embodiment of the present invention. More specifically, the step size for each frame is determined by using Equation 4 above, where a fixed parameter (

) Has a value of 10), and the signal threshold of each frame

It is an index indicating the watermark toughness when.

As shown in FIG. 9, the inserted watermark according to the preferred embodiment applicable to the present invention shows that the BER is about 5% or less for most signal attacks, thereby making it robust to the attack.

In operation S170, IFFT is applied to a signal of a frame having a watermark embedded therein and reconstructed into a signal of a spatial domain (S180). In addition, by applying the IDWT to the signal of the frequency band not selected in the step S120 to reconstruct the signal of the spatial domain (S190).

Thereafter, a watermark-embedded audio signal is generated using the signals reconstructed in steps S180 and S190 (S200). In more detail, IFFT is applied to the signal in the frequency domain where the watermark is inserted and reconstructed into a signal in the spatial domain.In step S110, IDWT is applied to the signal that is DWT-converted and the watermark is not inserted, Reconstruct the signal. In this way, the watermark-embedded audio signal is generated using each signal reconstructed as a signal in the spatial domain.

10 is a block diagram showing an audio watermark detection apparatus according to another embodiment of the present invention.

As shown in FIG. 10, in a preferred embodiment applicable to the present invention, an audio watermark detection apparatus applies maximum energy among a plurality of frequency bands separated by applying a discrete wavelet transform (DWT) to an audio signal having a watermark embedded therein. A second band selecting unit 400 for selecting a frequency band to have; And splitting the signal of the frequency band generated by applying an inverse discrete wavelet transform (IDWT) and a fast fourier transform (FFT) to the selected frequency band signal into a plurality of frames, and determining the Euclidean distance for each frame sample. It includes a watermark detection unit 500 for detecting a watermark by using.

The second band selector 400 selects a frequency band having the maximum energy among a plurality of frequency bands separated by applying DWT to the audio signal having the watermark embedded therein. Hereinafter, a detailed description will be given with reference to FIG. 11.

11 is a block diagram illustrating a second band selecting unit according to another embodiment of the present invention.

As illustrated in FIG. 11, in a preferred embodiment applicable to the present invention, the second band selecting unit 400 applies a DWT to a watermark-embedded audio signal and divides it into a plurality of frequency bands. 410; And a second selector 420 for selecting a frequency band having the maximum energy among the plurality of frequency bands.

The second DWT converter 410 separates the watermarked audio signal into a plurality of frequency bands as shown in FIG. 12 by applying DWT to the watermarked audio signal. In the case of FIG. 12, the watermarked audio signal is divided into six frequency bands, but the present invention is not limited thereto. The DWT transformation has a characteristic that there is no blocking phenomenon by transforming the entire signal and that both frequency information and position information can be expressed due to the characteristics of the transfer function. The DWT also allows for multiple resolution analysis. DWT having such characteristics is applied to various fields such as signal processing.

The second selector 420 selects a frequency band having the maximum energy among the plurality of frequency bands. In the preferred embodiment applicable to the present invention, since the watermark is inserted into the frequency band having the maximum energy for watermarking, which is robust to signal processing, the frequency band having the maximum energy can be selected as the watermark detection band.

The watermark detection unit 500 separates the signal of the frequency band generated by applying IDWT and FFT to the selected frequency band signal into a plurality of frames, and uses the Euclidean distance for the samples of each frame. Detect. Hereinafter, a detailed description will be given with reference to FIG. 13.

13 is a block diagram illustrating a watermark detection unit according to another embodiment of the present invention.

As shown in FIG. 13, in a preferred embodiment applicable to the present invention, the watermark detector 500 includes: a third IDWT converter 510 for reconstructing the band signal by applying IDWT to the signal of the selected band; A second frame converter 520 for applying an FFT to the reconstructed signal and separating the FFT-converted signal into a plurality of frames; A second step size determination unit (530) for determining a step size for each frame by using a frequency response size of each frame of the separated frame; A second threshold setting unit (535) for setting a signal threshold of each frame by using a step size corresponding to each frame; And a detector 540 for detecting the watermark using the step size and Euclidean distance corresponding to each frame.

The third IDWT converter 510 reconstructs the band signal by applying IDWT to the selected band signal. More specifically, before detecting a watermark for the selected frequency band, IDWT is applied to the signal and reconstructed into a signal of a spatial domain.

The second frame converter 520 converts the signal into a frequency domain by applying an FFT to the reconstructed signal. The present invention relates to an apparatus for detecting a watermark inserted in a QIM scheme in a frequency domain. In an embodiment applicable to the present invention, the second frame converter 520 applies an FFT to a signal reconstructed into a spatial domain. By converting the signal into a signal in the frequency domain, the watermark can be detected in the frequency domain.

In addition, the second frame converter 520 separates the FFT-converted signal into a plurality of frames. The present invention relates to a watermark embedding apparatus for inserting a watermark (1) bit into a sample of each of the separated frames, and to a watermark detecting apparatus for detecting the inserted watermark, in an embodiment applicable to the present invention. The second frame converter 520 may be divided into frame units including M (M is a natural number) samples without overlap by using a rectangular window.

The second step size determination unit 530 determines the adaptive step size for each of the separated frames using the frequency response size of each frame. In a preferred embodiment applicable to the present invention, the second step size determination unit 530 may determine the step size for each frame using Equation 7 below.

는 k번째 프레임의 스텝사이즈,

는 k번째 프레임의 주파수 응답 크기,

는 신호 전체에 대해 고정된 파라미터)(

Is the step size of the kth frame,

Is the magnitude of the frequency response of the kth frame,

Is a fixed parameter for the entire signal)

)는 각 프레임별 스텝사이즈의 크기를 조절하는 값으로, 오디오 워터마크 검출장치의 환경 정보 또는 워터마크가 삽입된 신호의 특성에 따라 상기 고정된 파라미터(

)의 크기는 변경될 수 있다. 본 발명에 대한 일 실시예에서는 상기 고정된 파라미터(

) 값으로 10을 적용하였으나, 이는 일 실시예로 본 발명은 이에 국한되지 않는다.
Fixed parameter of Equation (7)

) Is a value for adjusting the size of the step size for each frame, and the fixed parameter (

) Can be changed in size. In one embodiment of the present invention the fixed parameter (

Although 10 is applied as a value, the present invention is not limited thereto.

The second threshold setting unit 535 may set the signal threshold of each frame by using the step size corresponding to each frame. In an embodiment applicable to the present invention, the second threshold value setting unit 535 sets the signal threshold value of each frame to the first threshold value setting unit 235 of the audio watermark embedding apparatus corresponding to the audio watermark detection apparatus. It can be equal to the signal threshold of each frame of. When the watermark is inserted, the watermark is inserted only for samples having a predetermined frequency response or more. Accordingly, the second threshold setting unit 535 may detect the watermark only for samples having a frequency response size or larger. .

The detector 540 detects the watermark using the step size and Euclidean distance corresponding to each frame. In more detail, as shown in Equation 8, a method of determining a watermark value such that a difference value between a signal having a watermark embedded therein and a quantized signal by using a dither vector d (m) becomes the minimum value. Detect watermark through

가 신호 문턱값 이상의 주파수 응답 크기를 갖는 샘플에만 삽입되므로, 상기 검출부(540)는 상기 각 프레임별 샘플의 주파수 응답 크기가 각 프레임의 신호 문턱값 이상인 샘플에 대해서만 워터마크를 검출할 수 있다.
One bit watermark in the embodiment applicable to the present invention

Since is inserted only to a sample having a frequency response magnitude greater than or equal to a signal threshold, the detector 540 may detect a watermark only for a sample whose frequency response magnitude of the sample for each frame is greater than or equal to the signal threshold of each frame.

14 is a flowchart illustrating an audio watermark detection method according to another embodiment of the present invention.

As shown in FIG. 14, in a preferred embodiment applicable to the present invention, an audio watermark detection method includes: separating a plurality of frequency bands by applying DWT to an audio signal having a watermark embedded therein; Selecting a frequency band having the maximum energy among the plurality of frequency bands; Reconstructing the band signal by applying IDWT to the signal of the selected band; Applying an FFT to the reconstructed signal, and separating the FFT transformed signal into a plurality of frames; Determining a step size for each frame using the frequency response magnitude of each frame of the separated frame; Setting a signal threshold of each frame using the step size corresponding to each frame; And detecting the watermark using the step size and Euclidean distance corresponding to each frame.

The DWT is applied to the watermark-embedded audio signal to separate the watermark-embedded audio signal into a plurality of frequency bands as shown in FIG. 15 (S510). In the case of FIG. 15, an audio signal having a watermark embedded therein is divided into six frequency bands. However, the present invention is not limited thereto. The DWT transformation has a characteristic that there is no blocking phenomenon by transforming the entire signal and that both frequency information and position information can be expressed due to the characteristics of the transfer function. The DWT also allows for multiple resolution analysis. DWT having such characteristics is applied to various fields such as signal processing.

The frequency band having the maximum energy among the plurality of frequency bands separated in step S510 is selected as the watermark detection band (S520). In the preferred embodiment applicable to the present invention, since the watermark is inserted into the frequency band having the maximum energy for watermarking, which is robust to signal processing, the frequency band having the maximum energy can be selected as the watermark detection band.

The band signal is reconstructed by applying IDWT to the signal of the band selected in step S520 (S530). More specifically, before detecting a watermark for the selected frequency band, IDWT is applied to the signal and reconstructed into a signal of a spatial domain.

The FFT is applied to the reconstructed signal in step S530 to convert the signal into a frequency domain signal, and the FFT-converted signal is separated into a plurality of frames (S540).

The present invention relates to a method for detecting a watermark inserted by a QIM method in a frequency domain. In an embodiment applicable to the present invention, an FFT is applied to a signal of a spatial domain reconstructed in step S530 to convert into a signal of a frequency domain. By converting, the watermark can be detected in the frequency domain.

In addition, the FFT transformed signal is separated into a plurality of frames. The present invention relates to a watermark embedding method for inserting one (1) bit of watermark into a sample of each of the separated frames, and a watermark detection method for detecting the inserted watermark, in an embodiment applicable to the present invention. In step S540, the rectangular window is separated into frame units including M (M is a natural number) samples without overlap.

An adaptive step size is determined for each frame separated in step S540 using the frequency response size of each frame (S550). In the preferred embodiment applicable to the present invention, the step S550 determines the step size for each frame using Equation 9 below.

는 k번째 프레임의 스텝사이즈,

는 k번째 프레임의 주파수 응답 크기,

는 신호 전체에 대해 고정된 파라미터)(

Is the step size of the kth frame,

Is the magnitude of the frequency response of the kth frame,

Is a fixed parameter for the entire signal)

)는 각 프레임별 스텝사이즈의 크기를 조절하는 값으로, 오디오 워터마크 검출장치의 환경 정보 또는 워터마크가 삽입된 신호의 특성에 따라 상기 고정된 파라미터(

)의 크기는 변경될 수 있다. 본 발명에 대한 일 실시예에서는 상기 고정된 파라미터(

) 값으로 10을 적용하였으나 이는 일 실시예로 본 발명은 이에 국한되지 않는다.
Fixed parameter of Equation (9)

) Is a value for adjusting the size of the step size for each frame, and the fixed parameter (

) Can be changed in size. In one embodiment of the present invention the fixed parameter (

10 is applied as a value of), but the present invention is not limited thereto.

In an embodiment applicable to the present invention, the signal threshold of each frame may be set using the step size corresponding to each frame (S560). In this case, the signal threshold of each frame may be the same as the signal threshold of each frame in the audio watermark embedding method corresponding to the audio watermark detection method. When the watermark is inserted, the watermark is inserted only for a sample having a frequency response size or more. Therefore, the watermark may be detected only for a sample having a frequency response size or more.

The watermark is detected using the step size and Euclidean distance corresponding to each frame (S570). More specifically, as shown in Equation 10, a method of determining a watermark value such that a difference value between a signal having a watermark embedded therein and a quantized signal using the dither vector d (m) becomes the minimum value Detect watermark through

가 신호 문턱값 이상의 주파수 응답 크기를 갖는 샘플에만 삽입되므로, 상기 S570 단계에서 상기 각 프레임별 샘플의 주파수 응답 크기가 각 프레임의 신호 문턱값 이상인 샘플에 대해서만 워터마크를 검출할 수 있다.
One bit watermark in the embodiment applicable to the present invention

Since is inserted only to a sample having a frequency response size greater than or equal to a signal threshold value, the watermark may be detected only for a sample in which the frequency response size of the sample for each frame is greater than or equal to the signal threshold value of each frame.

So far I looked at the center of the preferred embodiment for the present invention. Those skilled in the art will appreciate that the present invention can be implemented in a modified form without departing from the essential features of the present invention. Therefore, the disclosed embodiments should be considered in an illustrative rather than a restrictive sense. The scope of the present invention is defined by the appended claims rather than by the foregoing description, and all differences within the scope of equivalents thereof should be construed as being included in the present invention.

100: first band selector 110: first DWT converter
120: first selection unit
200: watermark insertion unit 210: first IDWT conversion unit
220: first frame converter 230: first step size determiner
235: first threshold value setting unit 240: QIM unit
300: signal generator 310: first IFFT converter
320: second IDWT conversion unit 330: reconstruction unit
400: second band selector 410: second DWT converter
420: second selection unit
500: watermark detection unit 510: third IDWT conversion unit
520: second frame converter 530: second step size determiner
535: second threshold value setting unit 540: detection unit

Claims (18)

A first band selector configured to select a frequency band having a maximum energy among a plurality of frequency bands separated by applying a discrete wavelet transform (DWT) to an original audio signal;
A frequency band signal generated by applying an inverse discrete wavelet transform (IDWT) and a fast fourier transform (FFT) to the selected frequency band signal is divided into a plurality of frames, and a QIM (Qunatization Index Modulation) is applied to each frame sample. A watermark insertion unit for embedding the watermark in a manner; And
Inverse Fast Fourier Transform (IFFT) is applied to the frame signal with the watermark embedded therein, and an Inverse Discrete Wavelet Transform (IDWT) is applied to the signal of the frequency band not selected by the first band selector. Audio watermark embedding apparatus comprising a signal generator for generating an audio signal.
The method of claim 1,
The first band selection unit,
A first DWT converter which applies DWT to the original audio signal and divides the signal into a plurality of frequency bands; And
And a first selector which selects a frequency band having the maximum energy among the plurality of frequency bands.
The method of claim 1,
The watermark inserting unit,
A first IDWT converter configured to reconstruct the band signal by applying IDWT to the signal of the selected band;
A first frame converter applying an FFT to the reconstructed signal and separating the FFT-converted signal into a plurality of frames;
A first step size determination unit determining a step size for each frame by using a frequency response size of each frame of the separated frame; And
And a QIM unit for inserting a watermark into the samples for each frame by the QIM method by using the step size corresponding to each frame.
The method of claim 3, wherein
The watermark inserting unit,
The apparatus may further include a first threshold setting unit configured to set a signal threshold of each frame by using a step size corresponding to each frame.
The QIM unit,
An audio watermark embedding apparatus, characterized in that a watermark is inserted only for samples whose frequency response magnitude is equal to or greater than the signal threshold of each frame.
The method of claim 1,
Wherein the signal generator comprises:
A first IFFT converter configured to reconstruct a signal in a spatial domain by applying an IFFT to the frame signal having the watermark embedded therein;
A second IDWT converter configured to reconstruct a signal in a spatial domain by applying IDWT to a signal in a frequency band not selected by the first band selector; And
And a reconstruction unit configured to generate an audio signal having a watermark embedded therein using the signals reconstructed by the first IFFT converter and the second IDWT converter.
Selecting a frequency band having a maximum energy among a plurality of frequency bands separated by applying a discrete wavelet transform (DWT) to the original audio signal;
A frequency band signal generated by applying an inverse discrete wavelet transform (IDWT) and a fast fourier transform (FFT) to the selected frequency band signal is divided into a plurality of frames, and a QIM (Qunatization Index Modulation) is applied to each frame sample. Embedding the watermark in a manner; And
Inverse Fast Fourier Transform (IFFT) is applied to the frame signal embedded with the watermark, and an Inverse Discrete Wavelet Transform (IDWT) is applied to a signal of a frequency band not selected in the band selection step. Audio watermark embedding method comprising the step of generating a signal.
The method according to claim 6,
The step of selecting the frequency band,
Applying DWT to the original audio signal and separating the plurality of frequency bands; And
And selecting a frequency band having the maximum energy among the plurality of frequency bands.
The method according to claim 6,
Inserting the watermark,
Reconstructing the band signal by applying IDWT to the signal of the selected band;
Applying an FFT to the reconstructed signal, and separating the FFT transformed signal into a plurality of frames;
Determining a step size for each frame using the frequency response magnitude of each frame of the separated frame; And
And embedding a watermark in the QIM method in each sample for each frame using the step size corresponding to each frame.
The method of claim 8,
Inserting the watermark,
Setting a signal threshold value of each frame by using a step size corresponding to each frame;
Inserting a watermark in the QIM method,
An audio watermark embedding method, characterized in that a watermark is inserted only for samples whose frequency response magnitude is equal to or greater than the signal threshold of each frame.
The method according to claim 6,
The signal generation step of generating the audio signal embedded with the watermark,
Reconstructing a signal of a spatial domain by applying an IFFT to the frame signal having the watermark embedded therein;
Reconstructing a signal of a spatial domain by applying IDWT to a signal of an unselected frequency band in selecting the maximum energy frequency band; And
And generating a watermark-embedded audio signal by using the reconstructed signals through applying the IFFT and applying the IDWT.
A second band selector configured to select a frequency band having a maximum energy among a plurality of frequency bands separated by applying a discrete wavelet transform (DWT) to an audio signal having a watermark embedded therein; And
A frequency band signal generated by applying an inverse discrete wavelet transform (IDWT) and a fast fourier transform (FFT) to the selected frequency band signal is divided into a plurality of frames, and a Euclidean distance is used for each frame sample. And a watermark detection unit for detecting the watermark.
12. The method of claim 11,
The second band selection unit,
A second DWT converter which applies DWT to the audio signal into which the watermark is inserted and divides the signal into a plurality of frequency bands; And
And a second selector which selects a frequency band having the maximum energy among the plurality of frequency bands.
12. The method of claim 11,
The watermark detection unit,
A third IDWT converter configured to reconstruct the band signal by applying IDWT to the signal of the selected band;
A second frame converter configured to apply an FFT to the reconstructed signal and separate the FFT-converted signal into a plurality of frames;
A second step size determination unit determining a step size for each frame by using a frequency response size of each frame of the separated frame; And
And a detection unit for detecting the watermark by using the step size and Euclidean distance corresponding to each frame.
The method of claim 13,
The watermark detection unit,
A second threshold setting unit configured to set a signal threshold of each frame by using a step size corresponding to each frame,
Wherein:
An audio watermark detection apparatus, characterized in that the watermark is detected only for samples whose frequency response magnitude is equal to or greater than the signal threshold of each frame.
Selecting a frequency band having a maximum energy among a plurality of frequency bands separated by applying a discrete wavelet transform (DWT) to an audio signal having a watermark embedded therein; And
A frequency band signal generated by applying an inverse discrete wavelet transform (IDWT) and a fast fourier transform (FFT) to the selected frequency band signal is divided into a plurality of frames, and a Euclidean distance is used for each frame sample. And detecting the watermark.
16. The method of claim 15,
The step of selecting the frequency band,
Applying DWT to the audio signal having the watermark embedded therein, and separating the frequency into a plurality of frequency bands; And
And selecting a frequency band having the maximum energy among the plurality of frequency bands.
16. The method of claim 15,
Detecting the watermark,
Reconstructing the band signal by applying IDWT to the signal of the selected band;
Applying an FFT to the reconstructed signal, and separating the FFT transformed signal into a plurality of frames;
Determining a step size for each frame using the frequency response magnitude of each frame of the separated frame; And
And detecting the watermark by using the step size and Euclidean distance corresponding to each frame.
18. The method of claim 17,
Detecting the watermark,
Setting a signal threshold value of each frame by using a step size corresponding to each frame;
Detecting the watermark using the step size and Euclidean distance corresponding to each frame,
An audio watermark detection method, characterized in that the watermark is detected only for samples whose frequency response magnitude is equal to or greater than the signal threshold of each frame.

Images