KR100698253B1 - Method for hiding informational message in audio signals - Google Patents

Abstract

The present invention is to provide a method of concealing an information providing message in an audio signal for protecting a hidden transmission or storage of an information message and copying. The output data is computed by Fourier transform for each time interval using the sub-watermark parameter with the same value as. The resultant watermark signal is inserted into the calculated watermark signal during the two channels coinciding with the carrier frequency, and is output after being changed. To output the initial audio signal.

Watermark

Description

Method for hiding informational message in audio signals}

1 is a view showing a watermarking insertion method according to the present invention.

2 is a view showing a watermarking detection method according to the present invention.

* Explanation of symbols for main parts of the drawings

10, 10 ': watermark parameter register 20, 50: temporary calculation unit

30: watermark part 40, 60: output generation part

The present invention relates to a digital watermarking method, and more particularly to a method of concealing an information message in an audio signal.

In general, with the development of information and communication technology, various data forms are changed from analog forms to digital forms which are easily processed and have no deterioration in quality. And the importance of such applications in cyber magazines and audio.                         

Among them, audio signals stored in recording media such as cassette tapes or CDs, that is, music, can be enjoyed through MP3 files provided on the Internet, which has recently formed the whole world as a single area. It is now possible to trade copyrighted materials such as audio data produced in digital form in the cyber market.

However, various media such as audio data produced in a digital form are easy to store or convert information, and when placed in an open space such as the Internet, a large amount of copying is possible without restriction of space and requestor, especially after copying. It retains the same data as the original, so it is likely to be copied and distributed without the author's consent. As a result, the intellectual property rights of the original authors are not protected.

Due to the convenience, mass replication, and circulation of digital information, the infringement of intellectual property rights through virtual space is becoming more serious, and various security devices have been developed to protect the intellectual property rights of the original authors. have.

One of them is the encryption method of digital information, which encrypts the transmitted data and transmits it, and the receiving side restores the encrypted data through the input of a public key or a private key, thereby illegally allowing a third party not authorized during transmission. Even if it is received, it can't be restored.

However, in this case, since the data may be disclosed to a third party according to the intention of the user who received the data legally, the data itself cannot protect the intellectual property rights of the original author.

Therefore, in recent years, a watermarking method has been used to detect and claim data tampering and forgery as an invisible structure of digital data and use it for authentication and property protection.

The above watermarking provides additional information, such as rights holders and certifications, to the original data for intellectual property rights such as multimedia data and publications that can be widely distributed and distributed as available on the network. It is a technique to insert and protect intellectual property of data. It is divided into visible watermarking and invisible watermarking.

And the characteristics of watermarking are inaudibility, statistical invisibility, strong in transmission, filtering, resampling, compression, digital-analog and analog-digital conversion, data format conversion, and Is inserted directly into the data.

In this case, a method of inserting watermarking is provided in various ways such as a spatial method, a frequency domain method, and spread spectrum communication.

The spatial method is a method of using a small change in the screen pixel value YIQ as a watermark in terms of space, and it is easy to insert a watermark, but has a problem in that it is weak in image processing such as loss, compression, and filtering.

In addition, the frequency domain method converts digital data into an analog signal of frequency components, and uses the same transform method to perform watermarks through discrete cosine transform, fourier transform, wavelet transform, or the like. How to convert and insert. This frequency domain method has an advantage that the inserted watermark coefficients are distributed over the entire area of the data, making it difficult to delete the inserted watermark, but there is a problem in that image loss such as spots or dents occurs depending on the coefficient value.

Lastly, the spread spectrum communication method in which the frequency domain method, which has been recently used and applied, is basically a method of converting and processing data into a frequency form. The spread spectrum communication method analyzes frequencies through discrete cosine transforms and then displays watermarks on the parts with the highest N coefficients in the transform matrix, i.

The spread spectrum communication method is similar to the frequency domain method, but employs a Code Division Multiple Access (CDMA) technique that spreads the frequency of the watermark to be inserted into the original data. Although watermarking has a more powerful advantage for image processing, there is a problem that it is not perfect in the process of compressing data.

Accordingly, an object of the present invention is to provide a method of concealing an information message that is impossible to listen to or watch and that is strong in the actual transmission of a digital audio signal.

Another object of the present invention is to provide easier fabrication by implementing a simple circuit.

A feature of the method of concealing an information message in an audio signal according to the present invention for achieving the above object is a watermarking method of an audio signal, the number of bits of the initial audio data and a watermark-inserted information sequence externally A first step of calculating the output data by Fourier transform for each time interval using a sub-watermark parameter having a value equal to and a new Fourier to the calculated watermark signal for the two channels matching the carrier frequency. And a third step of outputting the watermark signal by inserting and converting the coefficient into a coefficient, and outputting the watermark signal by performing inverse Fourier transform for each time interval. To lose.

In this case, the watermark is composed of a plurality of sub-watermarks and has the same value as one information providing bit.

And the sub-watermark has a carrier frequency less than 100 Hz.

The second step is

에

on

과, 보상신호 z인

를 대입하여 이루어지는 것에 그 특징이 있다.

And the compensation signal z

It is characterized by the fact that it is substituted.

In the watermark detection method of an audio signal, a digital audio signal including a watermark calculated in the watermarking step is input as an initial signal, and the number of bits of the input sequence and the information sequence into which the watermark is inserted The output data are computed by Fourier transform for each time interval using sub-watermark parameters with the same value. And using the calculated data to separate coefficients corresponding to the Fourier series, and outputting an estimated sequence of information bits matching the sub-watermark.

Other objects, features and advantages of the invention will become apparent from the following detailed description of embodiments taken in conjunction with the accompanying drawings.

A preferred embodiment of a method of concealing an information message in an audio signal according to the present invention will be described with reference to the accompanying drawings.

1 is a view showing a watermarking insertion method according to the present invention, Figure 2 is a view showing a watermarking detection method according to the present invention.

Prior to the description of the drawings, first, the calculation formula and its terms are defined.

First, the initial audio digital signals of the two channels input from the outside are referred to as x ₁ (t) and x ₂ (t).

In order to insert the watermarking, Fourier transform is performed using the first audio digital signal during each time interval [t ₁ , t ₂ ] as shown in Equation 1 below.

In Equation 1, x _i , i = 1, 2 and x _0i , i = 1, 2 are conditions including all other conditions of the Fourier series.

At this time, the watermarking method shown in Fig. 1 inserts sub-watermarks corresponding to one information bit, respectively.

Therefore, by substituting the sub-watermark corresponding to the information input bit (1 or 0) input from the outside into the Fourier-transformed equation (1) as the initial signal, the output of the two channels in the time interval [t ₁ , t ₂ ] The signal is represented by the following equation (2).

이고, N 은 시간 간격[t₁, t₂]내의 포인트 번호로, y1 - y2 = K =

로 정의된다.In Equation 2

N is the point number in the time interval [t ₁ , t ₂ ], where y1-y2 = K =

Is defined as

는 정보 신호의 진폭을 나타낸다.At this time, the

Denotes the amplitude of the information signal.

는 정보 비트와 일치하는 값으로, 1 일 때는 정보 비트와 서브-워터마크가 일치하고, 제로(zero)일 때는

또는 바이스 베르사(vice versa)이다.And,

Is the value corresponding to the information bit. If 1, the information bit and sub-watermark match, and if it is zero,

Or vice versa.

, 그리고

. 로 나타낼 수 있다.In addition, the parameter of the sub-watermark is a time interval [t ₁ , t ₂ ], the cycle frequency as shown in Equation 2

, And

. It can be represented as.

를 다르게 표현하면 다음 수학식 3과 같다.And the identity y1-y2 = K =

In other words, the following equation (3) is obtained.

는 초기 신호에 삽입된 정보 워터마크 신호를 의미하며, z는 초기 신호의 최소 잡음으로 선택된 보상신호를 의미한다.Where above

Denotes an information watermark signal embedded in the initial signal, and z denotes a compensation signal selected as the minimum noise of the initial signal.

The optimal value of this compensation signal is represented by the following equation (4).

을 갖는 채널의 경우는

과 같이 나타낸다.Signal level equivalent to Equation 3 and Equation 4

For channels with

It is represented as

)은 고정되거나, 또는 특별한 PDF(Portable Document Format)을 통해서 확률적으로 선택된다.And the amplitude of the information signal (

) Is either fixed or stochastic through a special Portable Document Format (PDF).

Let us look in detail at the concealment method of the overall information message through the drawings.                     

로 정의할 수 있다.Referring to FIG. 1, the watermark parameter register 10 stores sub-watermark parameters. If the sub-watermark parameter is set to M,

Can be defined as

M is equal to the number of bits in the information sequence in which the watermark is embedded.

And even if the time intervals [t ₁ , t ₂ ] coincide in other information bits, they do not have empty intersections if the frequency intervals do not coincide.

)을 정의하여야 한다.Therefore, before starting watermarking, the amplitude of the information signal during each time interval [t ₁ , t ₂ ] (

) Must be defined.

The parameter of the wrong sub-watermark number is thus defined in the same way as the use according to the sub-watermark.

And all the defined parameters are stored in the watermark parameter register 10.

을 수학식 2를 통해 계산한다. As described above, when all parameters are stored in the watermark parameter register 10 and external audio data is input from the outside, the temporary calculation unit 20 stores the data stored in the watermark parameter register 10 and the initial audio data. The Fourier transform is calculated at each time interval using the data as shown in Equation 1,

Is calculated through Equation 2.

Subsequently, the watermark unit 30 changes the channel coefficient a _i in the Fourier series, that is, the Fourier coefficient composed of the information part and the correction part, and inserts the watermark signal into two channels through Equation 3 and Equation 4.

The output generator 40 detects the watermark signal output from the watermark unit 40 and performs inverse Fourier transform for each time interval to output the initial audio signal having the watermark embedded therein.

Accordingly, the watermarking according to the present invention proceeds with a change in amplitude and a pulse of the signal inserted at the carrier frequency.

As described above, a method of detecting a watermark in the initial audio signal having the watermark inserted therein will be described with reference to FIG. 2.

Referring to FIG. 2, a digital audio signal including a watermark calculated in the watermarking step described with reference to FIG. 1 is externally input as an initial signal.

At this time, the watermark parameter register 10 'has the same information as described in FIG.

을 수학식 2를 통해 계산한다. The temporary calculator 50 calculates a Fourier transform as shown in Equation 1 at each time interval by using the data stored in the watermark parameter register 10 'and an input initial signal, and output data of two channels.

Is calculated through Equation 2.

과 K(y1 - y2)를 이용하여 푸리에 시리즈에 일치하는 계수를 분리하여, 서브-워터마크와 일치된 정보 비트의 추정 시퀀스를 출력한다.Then, the output generator 60 is output from the temporary calculation unit 50

And K (y1-y2) are used to separate the coefficients corresponding to the Fourier series, and output an estimation sequence of information bits matching the sub-watermark.

In this case, the K value is determined through an information providing bit corresponding to the sub-watermark.                     

That is, if 1 or 0 of the informational bit and K value match, the hard input schema can be used. In the case of soft input schema, the K value uses a soft input channel recovery algorithm.

Accordingly, the watermarking and watermarker detection are performed in a circuit that is as simple as software.

As described above, the method of concealing the information providing message in the audio signal according to the present invention has the following effects.

First, it is strong in transmissions such as filtering, resampling, compression, digital-to-analog and analog-to-digital conversion, and data format conversion, especially in the actual transmission of digital audio signals.

Second, relatively simple fabrication is possible with simple circuit implementation.

Those skilled in the art will appreciate that various changes and modifications can be made without departing from the spirit of the present invention.

Therefore, the technical scope of the present invention should not be limited to the contents described in the embodiments, but should be defined by the claims.

Claims (10)

In the watermarking method of an audio signal, And a watermark of each sub-watermark is inserted into two channels of the audio signal with a value corresponding to one information providing bit. The method of claim 1, And the sub-water mark is inserted into an initial signal corresponding to a carrier frequency for the two channels for a new Fourier coefficient, with Fourier coefficients inserted into the sub-water mark. The method of claim 2, And when carrier frequencies other than the carrier frequencies used for the sub-water marks exist, the respective carrier frequencies are superimposed on each other. In the watermarking method of an audio signal, A first step of calculating output data by Fourier transform for each time interval by using externally inputted initial audio data and a sub-watermark parameter having a value equal to the number of bits of a watermark-embedded information sequence; A second step of outputting after inserting a new Fourier coefficient into the calculated watermark signal for two channels of the audio signal corresponding to a carrier frequency, and outputting the new Fourier coefficients; Detecting the output watermark signal and performing inverse Fourier transform for each time interval to output an initial audio signal having a watermark embedded therein; . The method of claim 4, wherein And said watermark consists of a plurality of sub-watermarks and has the same value as one informational bit. The method of claim 4, wherein And the Fourier coefficients comprise an information portion and a correction portion. The method of claim 6, And said information portion has a signal opposite to the correction portion. The method of claim 4, wherein the second step

on

and, Compensation signal z

Method of concealing an informational message in an audio signal, characterized in that the substitution is made. In the watermark detection method in an audio signal, A first step of inputting a digital audio signal including a watermark calculated in the watermarking step as an initial signal; A second step of calculating output data by Fourier transform for each time interval using a sub-watermark parameter having a value equal to the number of bits of the input initial signal and a watermark-embedded information sequence; And separating a coefficient corresponding to a Fourier series using the calculated data, and outputting an estimated sequence of information bits matched with a sub-watermark. Concealment method. delete

Images